As a lover of all things tasty, my mind often turns to the kinds of food and drink that I love. As someone who works with taxonomies and thesauri, I tend to try to classify them. Often, though, I don’t get very far with it because I start to get hungry. However, I just ate, so I don’t think it’ll be so much of an issue this time.

Classifying food is at once extraordinarily basic and mind-bogglingly complex, depending on how deep you want to go with it. At its simplest, you have the USDA food pyramid (or plate, depending on when you’re talking about). As a simple guide to make sure your kids are eating the right proportions, the USDA guide can be helpful, if problematic. As a way to look at meaningful relationships between food, though, it’s simply too narrow to be of any real use.

To see a much more complex food classification system, one can simply enter a grocery store.  In there, thousands of food items sit on shelves, organized in a very specific, scientifically driven way. This organization, though, is based on sales maximization, not organizational consistency. That’s how rice ends up in multiple places, with the cheap basic stuff with the other staples and nicely packaged, and pricier styles in International Foods, or somewhere similar. So while there are very good reasons for how the items in a grocery store are arranged, this isn’t the kind of organization that I mean.

I’m thinking of organization based on what food and drink is and how it is viewed by cooks and eaters, not on how to boost sales of the latest in frozen pot pies. However, this can get extremely complicated.

With thousands of various types of food and drink in this world, questions immediately arise that confuse the issue. We all know what bread is, and we all know what cake is. They use nearly all the same ingredients and the result is similar, if very distinct. If we’re building a taxonomy, are they distinct concepts? Is cake a type of bread, or is the fact that one is eaten mostly for dessert, while the other generally isn’t, a big enough difference to keep them separated? How about a box of macaroni and cheese? Obviously, a greater part of what’s in that box is pasta, but in the grocery store, it’s generally nowhere near an actual package of macaroni. Does that little bag of weird cheese powder in the box make it an entirely different product? It seems like a subset to me.

There are problems like this everywhere, which makes attempts at organization seem futile. Where do we even start? In taxonomic terms, the basically useless (for our purposes) food pyramid gives us a few broadest terms to work with. It’s woefully incomplete, but it’s a place to start. Australia did something a little like this with their Australian Health Survey Classification System, which was designed “to group similar foods and report trends in consumption by food category.” While it’s useful and quite interesting from a public health perspective, the near-700 line spreadsheet makes it indecipherable for use by your average eater.

Unless all we want is an organized but flat list of foods and beverages, it seems we must decide on the purpose of the classification, because nothing is going to be one-size-fits-all. There isn’t a comprehensive food taxonomy out there, at least that I know of, but there some really intriguing things that people have done with very specific kinds of classification.

In his book, On Food and Cooking: The Science and Lore of the Kitchen, food scientist and writer Harold McGee features a two-page table (which I cannot post here) that features, on the y-axis, the names of commonly used herbs, and on the x-axis, the chemical compounds that give each its distinct flavor. One look at the table reveals how much is shared between different herbs. Say you’re cooking something and need basil, but are surprisingly all out of it. McGee’s table can show you what other herbs contain the chemical or chemicals that you need to match that flavor. You might get some extra stuff in the dish that you didn’t need, but you will have the flavors that you want.

Then there’s chef Marc Powell, who built a food app that reads menus, turns them into XML documents, and tags them with taxonomy-based metadata for taste, texture, and other food characteristics. This metadata can then be used to do make recommendations for balancing the flavors of a dish, providing a list of ingredients to concoct possible dishes, or any number of possibilities.

I would absolutely love to use that tool; it’s exactly the sort of thing that I want, though for it to work the way I have in my head, I don’t think a simple taxonomy, no matter how large, would be enough, precisely because of the complications that I describe above. On the other hand, an ontology that relates ingredients to associated recipes could be extremely useful. If I could just open my refrigerator or pantry, search in the ontology for the ingredients that I have in there, and have it return possible dishes that use only what I have would change the game for me. With the Internet of Things coming closer and closer to reality for the masses, this doesn’t seem all that far-fetched.

All of this talk about food has given me quite an appetite, but at least I could complete the thought this time.

Daryl Loomis
Access Innovations