On the Origin of Spices

How humans ignored some plant defenses and became attracted to their taste and smell

A display of fragrant spices and herbs


Excerpted from Delicious: The Evolution of Flavor and How It Made Us Human by Rob Dunn and Monica Sanchez. Copyright © 2021 by Rob Dunn and Monica Sanchez. Reprinted by permission of Princeton University Press.

For the first three hundred million years of mammal evolution, our ancestors ate from among the species available to them. They chose some flavors over others—mammoths over howler monkeys, for instance.

Some of our ancestors were choosier than others, but none could choose flavors that did not occur in nature. Nor do any among them appear to have mixed flavors, except inadvertently in the warm bowls of their mouths. Cooking offered new flavors, but the possibilities were finite. A mammoth foot roasted over an open fire could be cooked so that the meat was more or less moist and the skin more or less crackly, and yet it was inescapably always a mammoth’s foot. A breakthrough in the story of flavor occurred when our ancestors began to add spices to cooked foods. In doing so, they took advantage both of the diverse chemicals found in plants and the human ability to learn to enjoy almost any aroma. They created new mixes of aromas and tastes and learned to like those mixes.

As far as is known, no species except humans has learned to make a food by mixing multiple ingredients together. Chimpanzees do not add peas to their meat, nor do they add spices. In addition, spicing food is not universal among humans. Some groups of people do not use spices. Cornell University anthropologist Allen R. Holmberg (1909–1966), in his 1950 classic book, Nomads of the Long Bow: The Sirionó of Eastern Bolivia (Smithsonian Institution Press), observed that the Sirionó did not use spices in their cooking. The traditional diets of other Amazonian hunter-gatherer societies, including the Yano-mamö, seem similarly spiceless. One of the few exceptions appears to be the use of the ashes of some plant species as a form of salt. Amazonians are not unique in this regard. Many groups of people appear to have traditionally used no, or only few, spices in their foods.

The large deflated spheres on the surface of this basil leaf contain defensive chemicals that are released when the leaf is bitten, torn, or crushed.

Here, we use the term “spices” broadly to include the parts of plants used in food not for their nutritional value, but for some other reason, typically in small doses, and in ways that feature their aromas and flavors. Some spices, those that are often called “herbs,” are the leaves of plants—peppermint, spearmint, oregano, basil, bay, and lemon-grass. On many of these leaves are small, plump spheres in which the plant stores chemicals; when the spheres burst like small bombs, as in our mouths or when we cut or tear the leaves, the chemicals rise explosively into the air. Many spices, such as mustard, cumin, or anise, are seeds. Others, including chili peppers, black peppers, lemons, and limes, are whole fruits. Meanwhile, garlic, onions, and their many relatives are bulbs; cloves are flower buds; and saffron is the female sex part of a crocus.

From this large collection of saffron flowers (Crocus sativus), only the three crimson stigmata with their red pistils in the center of each flower are separated by hand and dried. It takes over 70,000 saffron flowers to make one pound of saffron spice, making it one of the most expensive spices by weight.

Joe McUbed/Shutterstock
The use of spices is a deceptively simple act. You add a little something else to your pot, pan, or bowl and, in doing so, alter the flavor of the resulting food. But it is not really so simple. In most cases, you need a pot, pan, bowl, or at least some container in which to mix ingredients—though spices can also be rubbed on the surfaces of cooking meats. But here is the other problem: essentially every spice that we use is from a strong-smelling plant part, whether a bulb, a leaf, or a seed. Most of those plants evolved the ability to produce the chemicals that we perceive as strong smells to ward off enemies.

Hundreds of millions of years ago the first plants colonized land. Much later, when the first animals crept ashore, those plants were relatively undefended. To the first terrestrial herbivores, the Earth was an endless salad. But only for a relatively brief period of evolutionary time. Plant species able to evolve toxic leaves and reproductive parts, such as seeds, were far more likely to survive. And so, they did.

Eventually, most plant species evolved defenses. Some were physical. For example, bits of silica in grassland plants deter even the biggest mammalian herbivores. Silica gives those grasses a terrible mouthfeel; eating plants that contain lots of silica is like eating a bed of lettuce on which someone has sprinkled sand. Many defenses, though, were chemical. Plants evolved chemicals that punish herbivores with convulsions, vomiting, and death. Such defenses often played a dual role of both deterring herbivores and killing the plant’s pathogens. In response to these chemicals, individual animal species evolved countermeasures (as did some pathogens), including specialized abilities to break down some chemical defenses. Plants, in response, evolved new defenses. This back and forth war helped to engender much of the diversity of plants and plant-eating animals on Earth. And the back and forth is ongoing. Many species and varieties of thyme (Thymus) grow wild throughout the Mediterranean. Different varieties of thyme produce different defensive aromas. Often, the variety on one hill produces a different aroma from that on a nearby hill. These differences have been shown to be partially determined by which herbivore, or other enemy, is most abundant. Where sheep are rare but slugs common, the thyme variety that produces an aroma that deters slugs thrives. Where sheep are common, the variety of thyme that produces aromas that sheep dislike is most common. Similarly, thyme basil (Acinos suaveolens), which also grows in the Mediterranean, produces much less of its most aromatic compounds when growing in areas that are inaccessible to goats and sheep. It offers less warning when there is no one to warn. Based on these and other observations, some scientists have gone so far as to suggest that the aromas of plants in the Mediterranean and Middle East are an effect of the thousands of years of herbivory by goats and sheep. The species and varieties that are left are those that were most strongly defended. The most geographically widespread thyme in Europe today, for example, is the variety that has the strongest chemical deterrents.

The war between herbivores and plants never completely ended anywhere. It never will. Yet, our bodies have signed truces with many individual plant species and lineages. The truces are manifest in our mouths as bitter tastes. Animals, including our ancestors, evolved bitter taste receptors to warn them away from plants they were unable to detoxify. These receptors are a little different for each animal species, as a function of what they are, and are not, able to detoxify. Bitter taste receptors allowed animals an easy system for knowing what not to eat. We have a bitter taste receptor, for instance, that tells us to avoid plants with strychnine, another that tells us to avoid caffeine. Fifteen different compounds in hops trigger at least one of three human bitter taste receptors. In return, the plants offered something. They evolved to produce aromas that are themselves not toxic, but warn of the presence of toxins.

To use a spice is to ignore nature’s admonishments. We intentionally gather plants with high concentrations of defensive chemicals or warning aromas and add them to our food, typically in small doses. The chemicals associated with the bitter tastes of dandelions and dill, for example, are poisons. The fragrant aromas of garlic, mint, thyme, and dill are warnings of the existence of poisons. Eating such plants despite their warnings is a bold act. Yet, it is one to which we have grown numb. We are so accustomed to the flavors and aromas of spices that we do not consider the unusualness of consuming them. With spices, then, we need to explain two things—how it is that we so readily convince ourselves that spices are pleasing, and why did we begin to spice foods and to enjoy spiced foods.

The first question—how we learn to enjoy spices—is the easier one to answer. A fetus experiences the tastes and aromas of the foods its mother consumes. The chemicals from the food enter the amniotic fluid and travel to the nose of the fetus. The fetus can sniff at the tiny sea within which it floats. Fetuses appear to be predisposed to learn that the maternal aromas in which they swim are pleasurable. This is true even if the aromas are defensive compounds of plants. For example, when mother sheep eat garlic, their amniotic fluid smells of the defensive chemicals in garlic. Fetal sheep smell the aroma. Having been exposed to it, they prefer it once they are born. If the amniotic fluid of pregnant rats is injected with garlic, the babies of those rats, upon birth, involuntarily try to suckle when presented with garlic; they pucker pink lips and search for their mothers.

In studies of humans, experiments have been less invasive, but the results have been similar. In a study published in the December 2000 issue of Chemical Senses, cognitive psychologist Benoist Schaal and colleagues at the Center for Smell, Taste and Food Science, CNRS, Dijon, France, compared two groups of expectant mothers in the Alsace region of France. During the last ten days of their pregnancy, women in one group were given as many samples as they wanted of anise-flavored mints, cookies, and syrup. The other group was given no anise-flavored food and was asked to refrain from eating any. The researchers then tested whether the newborn babies to the two groups of women differed in their fondness for anethole, the compound that gives anise its aroma. Babies whose mothers did not eat anise tended to show faces indicative of displeasure when exposed to a highly diluted aroma of anethole. In contrast, babies whose mothers ate anise were more likely to turn their heads toward the anethole smell, stick their tongues out, and move their tongues, as if to lick their lips.

garlic mustard


In another study of humans, babies whose mothers had eaten garlic during pregnancy puckered their lips to suckle at the aroma of garlic. Similar effects of the womb’s flavors have more recently been shown for peas, green beans, and sulfurous cheeses, such as Camembert, Munster, or Époisses. Eight-month-old babies whose mothers ate peas, green beans, and other green vegetables during pregnancy showed a preference for the aroma associated with the smell of greenness (2-isobutyl-3-methoxypyr-azine). The eight-month-olds whose mothers ate sulfurous cheese when pregnant showed a preference for the smell of dimethyl sulfide (which is present both in sulfurous cheeses and garlic). Mothers who ate fish while breast-feeding tended to have babies that enjoyed the aroma of fish, or at least the compound associated with the aroma of fish, trimethylamine. Trimethylamine is found in both the amniotic fluid and breast milk of fish-eating mothers. These effects of aromas associated with amniotic fluids and breast-feeding appear to sometimes, though not always, persist into childhood and later life.

But when and why did some humans start to use spices in the first place? And what reason did they have to learn to love them?

Western Hackberry (Celtis reticulata) is used as a spice by indigenous people in the desert Southwest of the United States.

Here and there in the archaeological record one finds evidence of what might be, and also might not be, spice use. Hackberries (Celtis sp.), for example, have been found in a 60,000-year-old Neanderthal hearth at Dederiyeh Cave in Syria. The hackberries of the region, like those in North America, are somewhat unpleasant and not terribly satisfying to eat on their own. Indigenous people in the desert Southwest of the United States use similar hackberries as a spice. They add them to meat dishes while the dishes are cooking, as one might use peppercorns. We do not yet know if Neanderthals were using hackberries for the same purpose.

One of the oldest well-documented instances of the use of spices is surprisingly recent, from an archaeological site dated no older than 6,600 years ago. The evidence comes from a study conducted by archaeologists Hayley Saul and Oliver Craig (at the time, Saul’s advisor) at the University of York in the UK, and colleagues in Spain, Denmark, and Germany. The study, published in the August 2013 issue of PLOS ONE, considered a number of archaeological sites. But the most detailed work was done on a site in northern Germany from a time in which agricultural cultures were spreading north and hunter-gatherers were undergoing transitions in their food ways. The site, Neustadt, was first occupied by hunter-gatherers about 4600 BCE, and continued to be occupied for another eight hundred years as the hunter-gatherers transitioned to agriculture. The researchers studied the transition between hunter-gatherer lifestyles and agricultural lifestyles based on how the ceramics made there and the food eaten at the site changed over time. The earliest inhabitants were known for making large, ceramic vessels of a style called “Ertebølle,” and are referred to as the Ertebølle people. The later agricultural inhabitants made smaller ceramics of a type called “funnel beaker,” and are referred to as the Funnelbeaker people.

Europe with inset of study area and sites where pot residues were acquired; including the Near East and northern Africa, indicating early contexts where various spices have been recovered.

Saul, Craig, and colleagues were able to find Ertebølle ceramic vessels from the site that contained what archaeologists call “foodcrusts.” These foodcrusts were evidence, first and foremost, that ancient northern Europeans were not very good at doing the dishes. But they could also be used to study what these people were eating. The Ertebølle hunter-gatherer foodcrusts contained both meat and plants, whereas the later funnel beaker ceramics tended to be more specialized, containing either meat or plants. Saul discerned, using a variety of laboratory approaches, that the meat in the Ertebølle foodcrusts was from wild animals, roughly half marine and half terrestrial, perhaps, from fish and deer. Meanwhile, Saul found that while some of the plant material was starch of some staple (perhaps, according to Craig, hazelnuts and acorns), much of it was from the seeds of garlic mustard (Alliaria petiolata). Garlic mustard is not related to garlic or leeks, but is instead a garlicky member of the mustard clan. Saul and Craig and their colleagues suggest that the garlic mustard they found caked in pots was being used as a spice in an ancient stew. The researchers also found beeswax in some cooking vessels, possibly honey added to the recipe. Craig has speculated that the Ertebølle hunter-gatherers began using ceramics so they could make these sorts of dishes. We suspect that the culinary aesthetic that included spices, whether in cooking pots or other contexts, likely began for different reasons in different cultures and with different spices. Some spice uses may represent idiosyncratic, aesthetic manifestations of culture. Others may have begun as a kind of culinary preventative medicine. Retired Cornell University professor Paul Sherman has argued that spices might have first been used in order to kill off pathogens present in food or to ward off the establishment of pathogens in food left overnight or over a few days—or food inadvertently left on imperfectly cleaned vessels. Humans may have, especially, used plant parts that were good at preserving leftovers and had strong aromas. This use of plants as spices might have been an extension of the more ancient use of plants as medicine. Even today, many spices are used as both medicines and flavoring. For example, a plant called bitter leaf (Vernonia amygdalina) is used by some people both as a medicine and as a spice in cooked foods. It is added to the Nigerian meat stew egusi. Sherman’s hypothesis matches up well with what is known about how we learn to like and dislike aromas, both in utero and after birth.

Among those spices with antimicrobial properties, garlics and other alliums (such as onions and leeks) are relatively well-studied examples. They have a unique chemical arsenal that they rely on for defense. In garlic, this arsenal depends on two key compounds, alliin and alliinase. These compounds are stored in separate chambers in a garlic bulb and come into contact only once the bulb is damaged. Alliinase is an enzyme. When the bulb is bitten—whether by an insect, a rodent, or a human—the alliinase makes contact with the alliin and instantly converts it to allicin. Allicin gives garlic its pungent aroma. Onions are similar except that a second reaction causes the allicin-like compound in onions to be transformed once more and become one of a variety of chemicals known as “lacrimators” (tear makers). Onions are not the only allium to produce lacrimators; garlics do too. But onions produce more of the compounds than do other alliums. Once in our eyes (or those of a forest rodent), these lacrimators irritate nerve endings and further break down into sulfuric acid and other even more bothersome compounds.

The stew me-e puhadi is made from a recipe written in cuneiform on a 3,600-year-old tablet.


Despite their strong defenses against being eaten, garlics and other alliums are featured in recipes in many parts of the world. These plants are good candidates for spices that some have learned to love, perhaps because of their antimicrobial ability. But no one had ever done the experiment of making an ancient dish with and without its garlic to see what happened. We decided we would. Recently, we worked with students at two high schools in Raleigh, North Carolina, to cook a dish with and without alliums, called me-e puhadi. We then studied what grew in the stew after it was left out at room temperature for several days. The recipe is one of a series of recipes written in cuneiform on a 3,600-year-old tablet held in the Yale Babylonian Collection and published in Ancient Mesopotamia Speaks (2019, Yale University Press), a companion book to an exhibition by the same name at the Yale Peabody Museum of Natural History. Nearly all the recipes feature more than one allium. We chose me-e puhadi because it featured four different alliums: onions, shallots, garlic, and leeks. Similar garlicky stews are also likely to have been eaten long before the time of ancient Babylon, both in the region of Babylon and beyond.

The students made the me-e puhadi recipe. They created replicate batches with and without alliums. They then watched what happened. The version without alliums spoiled very quickly and smelled terrible; that with alliums stayed more or less unchanged for several days.

If spices were more generally associated with preserving food, we should expect to see some broader patterns too in where and how they are used. We would expect spice use to be common where conditions are hot and wet, and pathogens grow quickly. This prediction is easy enough to make, but harder to test well. Sherman and one of his students, Jennifer Billing, tried one approach. They compiled recipes from around the world and then compared the mean number of spices in different recipes. What they found is that the hotter a region is, the more kinds of spices are found in the average recipe, as they had predicted. However, there are other reasons this pattern might be expected. For example, it is possible that more kinds of plants with the potential to be used as spices can grow in warm, wet places. In theory, it should be possible to statistically distinguish these two explanations, but no studies have done so, as of yet.

At least some plants produce compounds that appear to help steer their fruits to one disperser rather than another. Such is the case with chilies. Birds have a receptor in their mouths that detects heat, but it is a tiny bit different from that of mammals, different enough that the capsaicin in chilies—which causes the burning sensation in mammals—does not trigger the same reaction in birds. Chilies without capsaicin tend to get gnawed on by rodents, who are not likely to carry their seeds very far. But chilies with capsaicin are avoided by rodents. Meanwhile, birds eat the fruits and disperse the seeds to other areas. As an added bonus, chilies with capsaicin are also better defended against fungi. Chilies with capsaicin are more likely to get widely dispersed and more likely to survive once they get there.

The donne’ sali chili plant grows wild on the Mariana Islands where weather conditions are hot and humid year round.


We think that as we conduct more studies we will find that spices have served multiple roles in human history and prehistory, just as the compounds in spices play multiple roles in nature. Once humans began to store food for longer periods and settle more permanently—before the origin of agriculture, but probably not much before—some spices may have been added to food to help keep that food safe. The subconscious learning that goes on in human noses and brains made it easy to learn to love spices that helped keep people safe. Some spices also added pleasure to food, which would have been a particular benefit where settlements were becoming larger and the best-tasting species were becoming scarcer. The health benefits, flavors, and pleasure of spices would all in-crease as crops were domesticated and settlements became larger—and more prone to food-borne diseases—at the same time that the average dish was more likely to depend on a relatively bland staple such as rice, cassava, corn, or wheat. Once spices became common, they could then be subject to all of the vagaries of history. Some would become expensive by virtue of their rarity. Others would come to be viewed as magical, sexual, or some complex mix of the two. But all of these spices are based on the chemicals associated with the struggle of plants for persistence; their chemistry, however we use it, is a chemistry of defense, war, and reproduction, a chemistry that is just beginning to be understood and yet rises up from nearly every dish we consume.--RD, MS


Recent Stories

The way they live, the food they eat, and the effect on us

A true but unlikely tale

Story and Photographs by William Rowan

Increasing day length on the early Earth boosted oxygen released by photosynthetic cyanobacteria.

Genomic evidence shows that Denisovans and modern humans may have overlapped in Wallacea.