The test has since been repeated many times, with similar results. What it confirms is that, as we have seen, all human babies, from Sweden to China, have a strong innate preference for sweetness and a dislike of bitterness and sourness. Basic tastes are not a question of memory—we are hardwired to think sweetness is wonderful and bitterness is scary. No one has to learn these simple tongue reactions. But flavor is another matter. Flavors—these memories generated backward through our nose—are
all
learned. What we think about flavor in all its myriad forms, from toasted cumin to sea bass, from parsley to spaghetti carbonara, is not fixed. Each of us will have a different bank of memories and feelings about these; and it exists from day one, if not before.
Taste buds appear at seven or eight weeks of gestation. Already, by thirteen to fifteen weeks, the taste buds are mature. A thirteen-week-old fetus weighs maybe an ounce, with no fat under the skin, no air in the
lungs. Yet already he or she cannot only swallow but
taste
, and these sips of fluid leave memories.
In 2000, some French scientists did a remarkable experiment showing that newborns arrive in the world with a memory of how their particular amniotic fluid tasted. The mothers studied came from the Alsace region, where strong-tasting anise sweets are a local delicacy. Some of the women had eaten anise regularly during pregnancy, but some had not. The babies were tested straight after birth and again four days later, having tasted nothing outside the womb but milk. When an anise odor was wafted in front of them, the babies born to anise-eaters showed a marked and “stable” preference for anise. They turned their heads toward the anise smell, sticking their tongues out with a licking gesture. They remembered it, and apparently, it pleased them.
Further experiments have confirmed that other strong flavors, such as garlic, can also find their way into amniotic fluid. In one study, women agreed to swallow garlic capsules forty-five minutes before they were due for an amniocentesis; when it was tested, their amniotic fluid smelled garlicky. Babies born to voracious garlic eaters will have been floating in a sac of garlic water for nine months. It has been shown that babies exposed to garlic before birth are more likely to enjoy garlic in food later on. Likewise, mice whose mothers had been fed on artificial sweeteners when pregnant had an exaggerated taste for sweetness. Pregnant rats fed on junk-food chow—including savory snacks, sweetened cereals, and chocolate-hazelnut spread—had babies who also selected these foods over regular rodent pellets, though the babies’ preference for junk was lessened if the mothers switched to a healthier diet during lactation.
The flavors our mothers ingest most regularly can become like mother’s milk to us. Julie Mennella and Gary Beauchamp, biopsychologists working at the Monell Chemical Senses Center in Philadelphia, have done a series of experiments on how flavor in utero and in breast milk leaves children with lasting memories and preferences for certain foods. One of their most celebrated studies, from 2001, involved carrot juice. The babies of a group of mothers who drank carrot juice during the last trimester of pregnancy and again during the first two months of breastfeeding were predisposed to like the flavor of carrot. When the babies
were weaned onto solid food, several months after the mothers stopped drinking the carrot juice, they showed a marked preference for cereal flavored with carrot juice over plain cereal flavored with water.
The early exposure of babies to flavor—both in utero and through milk—works as a kind of “imprinting,” as Gary Beauchamp puts it. We become emotionally attached to these early aromas. As we saw in Chapter 1 with the “flavor window,” younger babies are more open than older ones to new tastes. When it comes to weaning, this is an argument—as we saw in Chapter 1—for ignoring the advice on exclusive breastfeeding for six months and offering early, varied bites of vegetable purees between four and six months. When it comes to the pre-food stage, however, flavor may be one of the strongest arguments for mothers attempting to breastfeed, at least for the first few months, and eating as varied a diet as possible while doing so. Some psychologists suggest that instead of saying to mothers, “Breastfeed for the baby’s good,” health-care advisers should say, “Breastfeed for your own good,” because if you do, you are likely to have a baby who is less fussy to deal with in the early stages of eating. Then again, I’ve known children (not my own) who’ve gone from formula milk at four months to black olives and spinach tart at twelve months, so it doesn’t always follow.
It is curious that we talk so little about the flavor of formula, given that it is the main thing many babies taste during that crucial first year. Because any given brand of formula milk does not vary, it seems to have an even greater “imprinting” power than breast milk. Babies who cannot tolerate regular cow’s-milk formula are sometimes given a special “hydrolysate” formula, whose proteins have been broken down (hydrolyzed) to make them more digestible. To adult tastes, these formulas taste especially nasty, with a sour cheesy tang and a strange hay-like odor. Mennella and Beauchamp followed children who had been fed with two different hydrolysate formulas. Objectively, both of these milks tasted equally unpleasant. But to the infants, the particular formula they had been assigned—sour or not—taught them how food should taste. When the two brands were switched round, the infants drank less: they preferred their own bad-tasting formula to the other one. More strikingly still, children aged four and five who had been fed on these sour-tasting hydrolysates as babies showed
more positive feelings about sour tastes and smells than children who had been fed on breast milk or regular formula. This is vivid proof that
anything
can start to taste good if you have enough positive memories of being fed it by a parent. The obvious implication is that formula-fed babies would benefit from having their milk flavored with vegetables.
Formula can never match the myriad benefits of breast milk, which range from lessening the risk of eczema and ear infections to reducing the likelihood of type 2 diabetes later in life to promoting healthy gut microbiota. But in the developed world, as we’ve seen, most mothers are unable or unwilling to breastfeed exclusively for the first six months. With each of my own babies, I gave it up for one reason or another (illness, work, bereavement, and a child with feeding difficulties) at three months. During the wait until one year, when they were old enough for regular whole cow’s milk, I’d have been glad to buy formula that was mildly scented with a range of green vegetables, just enough to give them a memory of spinach when the time came for them to try them for real.
Instead, in many countries, formula milk has been flavored, if at all, with vanillin, the artificial vanilla flavor that goes into industrially produced sweet foods, from ice cream to biscuits to cake. Vanilla milk has a long history. Back in 1940, the head nurse of the children’s hospital in Philadelphia recommended tempting reluctant feeders with three drops of vanilla essence in each bottle. Internet forums suggest that there are still many desperate parents who resort to vanilla extract when a baby rejects the bottle.
Since 1981, international food standards (the Codex Alimentarius of the World Health Organization) have stated that no flavorings should be added to infant formulas aimed at newborns. But vanillin is still a key ingredient in many of the “toddler milks” marketed for children aged one and over. In China, vanillin is prohibited in infant formula, but it continues to be illegally added by many manufacturers. In 2014, a team of chemical analysts found vanillin in four out of twenty samples of infant formula randomly purchased from supermarkets in the city of Wenzhou.
Of all the flavors you could think of with which to “imprint” a child, vanilla is possibly the least useful from a health standpoint (except, perhaps, chocolate: in 2010, the American company Mead Johnson withdrew its “premium” chocolate-flavored Enfagrow toddler milk amid com
plaints
from leading nutritional scientist Marion Nestle that it was training kids to “like candy”). The effects of vanilla milk are lasting. In 1999, some researchers in Germany tested the effects of the vanilla that had been in German “bottle milk” during the postwar years. The scientists asked 133 people to try two different ketchups, one of which was straight-up tomato ketchup and one of which, bizarrely, had been flavored with vanillin. (The reason the researchers chose ketchup was precisely that it is not normally associated with vanilla.) Of the respondents, the majority of those who had been breastfed had a preference for the pure ketchup, while the majority of those who had been reared on vanilla formula preferred the strange vanilla ketchup. Their baby milk had brainwashed these unfortunate people into thinking that vanilla made everything taste better.
Clearly, spinach milk would be a better plan, assuming it could be made safe for tiny infants. It will probably never take off, though. Over time, the odds are that babies would accept it and even prefer it, just as the hydrolysate babies with their bad-tasting formula think that milk is
meant
to taste sourish and cheesy. It’s the parents who would find vegetable milk hard to accept. We want our babies to have milk that corresponds to our own memories of childhood. Manufacturers know that you can only sell baby food by making it appealing to adults, which is why baby rusks are sometimes sweeter than doughnuts and why, for decades, until it was banned, jars of baby mush came seasoned with monosodium glutamate (MSG), to give it a more savory taste. When vanilla is found in baby foods, it has been put there to attract not the children themselves—who, as we’ve seen, can become emotionally bonded to flavors that are strange, sour, or strong, given the right memories—but to please adults. The babies are not the ones who buy the food. It is the grown-ups’ memories that the food companies are trying to appeal to. As they warm the sterilized bottle, parents sniff their baby’s milk; or maybe take a tiny sip. It is they, not the babies, who have memories of how childhood milk
ought
to taste: creamy and sweet, like milk left behind in the cereal bowl.
Do you remember your first passion fruit, your first avo
cado, your first Thai green curry? Such flavor memories can seem
inconsequential, the stuff of gastronomes. “Ah, yes, it was in Marseille in 1987 that I first tasted an authentic bouillabaisse.”
Yet, from the perspective of neuroscience, food memories are not something slight. Registering different flavors is one of the main ways that our bodies interact with the world around us. Amazingly enough, the human olfactory bulb is the only part of the central nervous system that is directly exposed to our environment, through the nasal cavity. Our other senses—sight, sound, and touch—need to travel on a complicated journey via nerves along the spinal cord up to the brain. Smell and flavor, by contrast, surge directly from plate to nose to brain.
Conventional wisdom used to be that humans have rather a weak sense of smell, compared to that of other animals: dogs, say (witness the fact that we don’t have sniffer humans at airports). But recent research suggests the contrary. We may not have a bloodhound’s ability to track a scent, but our olfactory discernment is second to none. We can detect a drop of Worcestershire sauce in a glass of tomato juice, or the scent of fear in another person’s sweat.
When I say that we discern smells and flavors, what I should really say is that we create them. Flavor is not actually
in
food, any more than redness is in a rose or yellow is in the sun. It is a fabrication of our brains, and for each taste, we create a mental “flavor image,” in the same way that we develop a memory bank of the faces of people we know. The difference is that whereas faces fade when you haven’t seen them in a while, flavors and smells have a way of lodging themselves indelibly. What you taste as a child is still there in your adult brain, even if you haven’t thought of it for years. The Norwegian Trygg Engen, the “founding father” of the study of smell and memory, characterized our sense of smell as “
a system designed not to forget
.”
In 1991, the biologists Richard Axel and Linda Buck discovered that olfactory receptors—cells in the nose that detect odor molecules—make up the largest single family in the human genome. Out of around 19,000 genes, Axel and Buck found, nearly 1,000—5 percent—are olfactory receptors. Their research finally unlocked some of the mystery of how humans can remember and discriminate among so many flavors and smells (and, thirteen years later, won them a Nobel Prize).
What makes the human system of olfaction so sophisticated is not just the receptors themselves, but the way they interact with our large brains. Each receptor cell is fairly specialized: it can detect only a small number of substances. But when you smell or taste something—a loaf of freshly baked bread, or lemon zest sprinkled over a stew—the receptors send messages to the olfactory bulb in the brain. Here, each flavor becomes encoded in its own particular pattern in a part of the olfactory bulb called
glomeruli
. A glomerulus has been described as a “detection point par excellence.” Each and every time you taste or smell something, the relevant glomerulus will take a snapshot of it. These snapshots show up in the brain as patterns, like a map.
Humans can distinguish around 10,000 separate smells, estimates Linda Buck. We walk in the house and instantly know that someone is cooking roast chicken for supper, and that they decided to stuff it with rosemary instead of thyme. Our olfactory systems have an immense power to detect flavors. Molecules that look nearly identical, to a specialist chemist in a lab, will be easily distinguished by an ordinary person who smells them. Our brains will also interpret the same chemical in radically different ways depending on how concentrated it is. Buck and colleagues note that a “striking example is a substance called thioterpineol, whose odor is described as ‘tropical fruit’ at a low concentration, as ‘grapefruit’ at a higher concentration, and as ‘stench’ at a still higher concentration.”