How—and Why—Did Cultural Tastes for Spicy Food Develop?
Moshe Hoffman and Erez Yoeli Explore the Functions of Seemingly Random Preferences
In this excerpt from chapter two of our upcoming book, Hidden Games: The Surprising Power of Game Theory to Explain Irrational Human Behavior, we discuss how peoples’ taste for spice developed to serve a hidden function: mitigating the risk of foodborne illness.
How come Indians developed a taste for very spicy cuisine? Why didn’t Eastern Europe’s Ashkenazi Jews do the same? In 1998, Jennifer Billing and Paul Sherman proposed that the answer lay in spices’ ability to inhibit and kill bacteria that cause food to spoil. This ability, they figured, was most useful in hot climates, so this is where we’d expect to find people who have learned to love spice.
To test their theory, Billing and Sherman began by documenting that spices do indeed have the ability to inhibit and kill bacteria. They did this by considering the thirty bacteria that most often cause food poisoning and then combining dozens and dozens of studies that tested whether the presence of a spice, or its active ingredients, slowed the growth of—or outright killed—one of these bacteria. Some spices, like allspice and oregano, and the root vegetables garlic and onion inhibit the growth of all thirty types. Many, like bay leaf, mint, coriander, and nutmeg, inhibit half to three-quarters of the bacteria.
Some, like black pepper, and the citric fruits lemon and lime only inhibit a small number of bacteria on their own but amplify the effect of other spices (black pepper by increasing the bioavailability of other spices’ active ingredients, thus increasing the rate at which they are absorbed by bacteria; lemon and lime by breaking down the bacteria’s cell walls, making the bacteria more susceptible to the active ingredients in spices).
Then, Billing and Sherman compiled hundreds of recipes from dozens of cultures around the world and indexed the types of spices and quantities used. Sure enough: “As mean annual temperature… increases, the proportion of recipes containing spices, number of spices per recipe, total number of spices used, and use of the most potent antibacterial spices all increase.” They then ruled out one alternative theory after another. It is natural to ask whether spices are simply used in hotter places because that’s where they grow. But, as the existence of a bustling and ancient spice trade suggests, spices are used in many more places than they’re grown.
Allspice is a nice example: it is used in ten times as many countries as it is grown. Moreover, it turns out spices aren’t even, on average, more commonly grown in hotter climates. In many hot places, few spices grow at all, but they’re still widely used. Wait, maybe spices can help people stay cool by causing them to sweat? Indeed, the capsaicin in chili peppers has this effect. But most spices, like oregano, mint, and cinnamon, do not. Maybe spices contain micronutrients that are somehow more useful in hotter places? Nope. Locally available vegetables and meat contain much higher concentrations of critical micronutrients.
So, back in their home countries, Indian mothers gradually trained their babies to tolerate and then love spicy food (by mixing adult food with yogurt and gradually reducing the amount of yogurt), while Ashkenazi mothers stuck to salt. When they came to America, they brought their cuisines with them.
As with the Fijians and Inuit, these Indian and Ashkenazi mothers need not have known why they ate the way they did. Most would have no idea and might respond: “That’s just how we’ve always done it,” or “It tastes good this way.” Nor would they have been aware of the learning processes that worked their magic over many generations, to shape their behaviors—by shaping their tastes.
Why might learning shape beliefs and tastes, rather than just shaping the desired behaviors directly? One possibility is the one we raised earlier: it’s simply an effective way to get us to act. If you like spicy food, you’ll eat spicy food. Another possibility, though, is that internalization ensures the functional behavior is taken even when its function is not understood. Like overimitation. If Indians like the taste of black pepper in their curry, they won’t be tempted to remove it, even if they don’t realize that it’s there because it amplifies the effect of cumin, coriander, turmeric, and capsicum. If Europeans had had a chance to develop a taste for nixtamalized corn, they’d have avoided pellagra.
Visit an Indian American family during Thanksgiving, for instance, and you might find that the turkey has been lightly curried or marinated overnight in yogurt and tandoori spices before being roasted whole. The cranberry sauce might look from afar like any other but reveal itself to have a touch of heat from green chilis and may be spotted with earthy, toasted cumin seeds. The gravy might have the golden yellow glow and unmistakable fragrance of saffron and perhaps a bit more black pepper than usual.
Nor is it just Indian Americans. Visit a Mexican American household, and you might find the turkey served alongside pozole and other traditional dishes from south of the border. One Mexican American friend recounts the time her aunt was charged with making green bean casserole. Her aunt looked up the recipe, decided it made no sense, and showed up with green beans and chili, instead. The entire family was thrilled.
Of course, neither the lox at Russ & Daughters nor the Thanksgiving turkey and green beans are at serious risk of having spoiled. The American food supply benefits from abundant refrigeration, as well as industrial food safety practices and regulations that minimize such concerns. Rather, immigrants from warm locales skip the lox or spice their turkey and fixings because they like the spice. They have, figuratively and literally, developed a taste for it—a taste that developed in a culture that needed to protect against foodborne illness but that persists even though they now live in a culture with no such need.
We will call such effects lags. Lags are like vestigial traits— human tails or the whale’s hand. They once had a purpose, but that purpose is no longer relevant. Lags occur because, like evolution, learning isn’t instantaneous. Just as it takes time for an individual to learn to like spicy food and for a culture’s cuisine to become spicy, it takes time for these things to be unlearned, or, as animal behaviorists would say, for extinction to occur.
Sometimes, we also run into situations where a belief or taste still has a relevant purpose in one context, but it persists even when taken out of this context. We call such effects spillovers. When your dog humps your leg, that’s a spillover. After all, humping some things has a very useful purpose, but a leg? Not so much. Spillovers occur because learning, like evolution, involves generalization— applying what we’ve learned in one context to other similar contexts—and while we’re pretty good at generalizing, we’re not perfect at it. Sometimes, we overgeneralize, and this is especially true when learning shapes our internalized beliefs and tastes.
Spillovers will be particularly helpful for understanding the results of psychology and economics laboratory experiments (that are, by design, performed in highly controlled laboratory settings). The laboratory environment is extremely useful: by tricking our tastes and beliefs into spilling over from day-to-day life into our experimental settings, we can document their many, fascinating quirks. (Nowadays, the word laboratory is in fact a bit of a misnomer. In many studies, the “laboratory” is Amazon Mechanical Turk, where people from far and wide can make a few extra bucks by logging in and anonymously answering surveys in their pajamas.)
Consider the following classic experiment. Subjects are partnered up in pairs to play the ultimatum game. One subject is given a few dollars—say ten dollars—and chooses how much to share with her partner. Her partner then chooses whether to accept the offer. If he accepts, they each keep their share. If he rejects, neither gets any money. Then, the subjects go their separate ways, never to meet again (literally—remember, these are strangers, on the internet, who have no way of identifying each other or even of communicating).
In the ultimatum game, it is very common to see fifty-fifty splits. Pretty intuitive. It is also common to see that when the first subject offers too little, say one or two dollars, her partner rejects the offer. Again, intuitive. And it makes sense: after all, our sense of justice is designed to prevent people from walking all over us, and when they do, we must strike back. The experiment is working beautifully: our sense of justice is nicely spilling over to this silly little game.
But wait. The experimenter has carefully designed the experiment so that it is anonymous. She has even made it so that she herself cannot identify the subjects. Nobody—not even the experimenter—will find out if certain subjects are pushovers and accept low offers. What’s the use in being outraged? Subjects should, if they are behaving rationally, bite the bullet and accept the low offer, which is better than nothing. So, does this mean people’s sense of justice does not have the function we claimed— of keeping others from trampling them?
We don’t think so. That would be like concluding that because dogs sometimes hump people’s legs, their sex drive didn’t evolve for reproduction. Rather, the correct interpretation, we think, is just that our sense of justice isn’t perfectly tuned to the specifics of every situation—not surprising given how unusual the fully anonymous laboratory setting is and how learning processes lag and spillover.
Our tastes are often taken as given, or seen as random–even irrational. But, as this excerpt illustrates, they often turn out to have a function. You just need to dig a bit to discover it.
Adapted from Hidden Games: The Surprising Power of Game Theory to Explain Irrational Human Behavior by Moshe Hoffman and Erez Yoeli. Copyright © 2022 by Moshe Hoffman and Erez Yoeli. Used with permission from Basic Books.