Desire

The Latin verb desiderare is built from de- and sidus — sidus meaning “heavenly body, star, constellation.” The literal sense is from the stars, or away from the stars. Cicero used the word of soldiers who desiderant — long for, miss — a commander who had not returned, or comrades who had not come back from battle. The Romans named the longing of those who could no longer see the constellations they had been promised.

Its companion word is considerare — con- plus sidus — to look at the stars together, carefully. To consider is collaborative stargazing. To desire is to have lost the sight, and to ache for it. Two acts of attending to the night sky, encoded in the prefix alone, and English keeps the two words sounding unrelated.

We have been desiring — etymologically, missing the stars — for two thousand years without noticing.

In 1954, in Donald Hebb’s neuropsychology lab at McGill, a thirty-two-year-old postdoctoral fellow named James Olds was implanting electrodes into rat brains. He was trying to reach the reticular formation in the brainstem — a structure recently implicated in arousal. The electrode missed. It came to rest in the septal area, a forebrain structure near what is now called the nucleus accumbens.

Olds and his supervisor Peter Milner — Cambridge-trained, careful with method, the husband of Brenda Milner who would soon do the H.M. memory work next door — put the rat in an open field and watched what happened when they delivered brief pulses of current through the misplaced wire. The rat went back to the spot where the stimulation had occurred. They rigged a Skinner box with a lever. The rat pressed it. It pressed it again. It pressed at six to thirty times per minute, for twelve hours straight, ignoring food, water, and a receptive female placed in the cage. In later optimized experiments, rats reached thousands of presses per hour. They would self-stimulate until they collapsed.

Kent Berridge and Terry Robinson at the University of Michigan, beginning in the mid-1980s, did a series of experiments that took the reward circuit apart. Their methodology was simple: rate the liking facial expressions of rats — tongue protrusions, paw-licks, the same expressions human infants make at sweet flavors — and separately measure how hard the rats worked to get the substance.

The dissociation was clean and unsettling. Rats whose dopamine systems had been destroyed by 6-OHDA lesions would starve in the presence of palatable food. They no longer wanted it. But if the food was placed in their mouths, they showed the same liking expressions as before. The pleasure system was intact. The wanting system was dead.

Hyper-stimulating the dopamine system reversed the picture. The animals worked harder to get the reward — pressing levers, traversing mazes — and showed no change in how much they appeared to enjoy it when they got it. Berridge and Robinson formalized this in 1993 as the incentive sensitization theory of addiction. Addiction, on this account, is not increased pleasure from the drug. Tolerance often reduces the pleasure. Addiction is pathologically amplified wanting of something the addict no longer enjoys.

Plato’s Symposium stages a drinking party at the house of the poet Agathon in 416 BCE. Seven men give competing speeches on the nature of Eros. Socrates’ speech is the longest, and he says he is not the author of it — he is relaying what was taught to him long ago by Diotima, a priestess from Mantinea.

Diotima’s central image is genealogical. At the birthday feast of Aphrodite, the god Poros — Resource, the son of Metis (cunning intelligence) — drank too much nectar and fell asleep in the garden. Penia — Poverty — slipped in, lay with him while he slept, and conceived Eros that night.

Eros, Diotima says, is therefore not a god but a daimon — intermediate between mortal and immortal, between full and empty, between knowing and not knowing. From his father he inherits the longing for what is precious. From his mother he inherits the structural condition of not having it. He is the child between them — always poor and far from delicate, hard and weather-beaten, sleeping in doorways; but also bold, enterprising, a hunter, full of resources. A god who only had Resource would be Zeus, full and satisfied. A god who only had Poverty would be a beggar at the door. The child is the structure between them.

Two and a half millennia later, every model of motivation that tries to explain why a satisfied human keeps wanting — every economic utility function, every prediction-error neuron, every theory of the hedonic treadmill — is a footnote to a story a priestess told a drunk philosopher about a god conceived in a garden.

Augustine wrote the Confessions in Hippo Regius around 397–400 CE. He had been a Manichee, a teacher of Latin rhetoric in Milan, and the unmarried partner of a concubine for fifteen years; she bore him a son, Adeodatus — “given by God.” Augustine had wanted, for most of those years, to live differently. In Book VIII, chapter 7, he records the prayer of his younger self with the irony of his older self looking back.

The Latin is shorter than the English: nine words, syntactically clean, asking God to fix what the petitioner does not yet want fixed. Augustine in his fifties is laughing at Augustine in his twenties. The prayer is a confession of how desire actually works — we petition for a future self who does not yet need to be us.

Later in the same book, after the scene under the fig tree where he hears a child’s voice singing tolle, lege — take and read — Augustine opens Paul’s letter to the Romans at a passage on putting off the works of the flesh, and the conversion is complete. He never returns to the concubine. He becomes Bishop of Hippo. He writes, earlier in the same volume, the line that names what the entire book is about: Inquietum est cor nostrum, donec requiescat in te — our heart is restless, until it rests in thee.

René Girard, French historian and literary critic, taught at Stanford from 1981 until his retirement in 1995. Sometime in the late 1980s, an undergraduate philosophy student named Peter Thiel took his seminar. Girard’s central argument, developed in Mensonge romantique et vérité romanesque (1961) — translated as Deceit, Desire, and the Novel — is simple and difficult. Human desire is not autonomous. We do not desire from ourselves. We desire what others desire, borrowing the wanting from models who confer value on objects by their own wanting.

Don Quixote reads books about the fictional knight Amadís of Gaul and then wants what Amadís wanted: glory, a horse, a beloved. The model is fictional, dead four centuries; the desire is real. Cervantes is the first novelist to put the borrowed structure of desire on Page 1 of the modern novel. Madame Bovary reads romantic novels in adolescence and then seeks in real lovers the heroes of the books. Flaubert builds the entire novel around the mediated structure. Anna Karenina, Proust’s narrator with Albertine, Heathcliff with Catherine — every great novel about desire, on Girard’s reading, is a story about a person who borrowed the wanting from somewhere.

When the model is far enough away — fictional, dead, of a different class — there is no rivalry. Girard called this external mediation. When the model is close enough to compete with — a colleague, a friend, a sibling — the desire becomes mutual and the object recedes; the rivalry is now the thing. This is internal mediation, and it is the structural source of nearly all human conflict.

Edward Bernays was Sigmund Freud’s nephew on both sides — his mother was Freud’s sister Anna; his father had married Freud’s wife Martha’s sister. The double family link mattered: Bernays did not just read his uncle. He absorbed psychoanalysis as something his uncle was working out at the breakfast table.

In 1928, George Washington Hill, president of the American Tobacco Company, hired Bernays to break the taboo against women smoking in public. Lucky Strike’s market would double if women smoked. Bernays consulted the psychoanalyst A. A. Brill, who suggested that cigarettes, in the unconscious, were “torches of freedom” — phallic symbols of emancipation.

On the morning of March 31, 1929, during the Easter Parade on Fifth Avenue, Bernays’ secretary Bertha Hunt and a group of recruited debutantes lit Lucky Strikes simultaneously and walked the parade route smoking. The story ran in the New York Times and across the country. Bernays had tipped off the photographers. No connection to American Tobacco was disclosed.

In 1997, Wolfram Schultz (then at Fribourg, later Cambridge), Peter Dayan (Gatsby Computational Neuroscience Unit, London), and P. Read Montague (Baylor) published a paper in Science titled “A Neural Substrate of Prediction and Reward.” They had recorded from single dopamine neurons in the midbrain of awake macaques while the monkeys learned to associate cues with juice rewards. The neurons did not fire in proportion to the juice. They fired in proportion to the prediction error — the difference between expected and received reward.

Unexpected reward: burst of firing. Reward predicted and delivered as expected: nothing. Reward predicted but withheld: a pause in firing, a negative blip below baseline.

In the same decade, Richard Sutton and Andrew Barto at Massachusetts were formalizing temporal difference learning — a reinforcement learning algorithm that updates predictions of future reward by the size of the prediction error. The algorithm had been published in computer science. Schultz’s data fit it neuron-for-neuron. Two disciplines had independently discovered the same algorithm; one had implemented it in monkey midbrain, the other in lines of code. Schultz, Dayan, and Montague’s paper has been cited over nine thousand times.

Denis de Rougemont published L’Amour et l’Occident — translated as Love in the Western World — in 1939. His central argument: Western romantic love is structured around impediment. The lovers’ passion is intensified by, and requires, obstacles — social, marital, geographical. Without the obstacle, the passion is not the same kind of thing. Tristan and Iseult drink the potion that binds them but the form of their story is the series of separations and impossibilities. When the lovers finally get to be together, the story ends with their deaths.

Anna Karenina runs the experiment in the form of a novel, sixty-two years before de Rougemont names what Tolstoy is doing. Anna leaves Karenin for Vronsky. The affair is sustained by the social impossibility of the union — St. Petersburg society opens and closes against her, the law refuses divorce, the child she has with Vronsky cannot inherit. The lovers go to Italy together. In a rented villa with paintings on the walls, in the absence of the impediments, Vronsky takes up painting; Anna reads English novels. The relationship begins to die.

It does not die because the passion was false. It dies because the passion was structured by the impediments. Vronsky becomes restless and goes to St. Petersburg without her. Anna begins to suspect — correctly — that his commitment is wavering. She follows him; the social system reasserts itself; the suffering returns. The novel ends at the train station.

A word that literally means “from the stars,” in a Latin verb Cicero used of soldiers longing for an absent comrade. A circuit in the rat brain discovered in 1954 by an electrode in the wrong place. A pleasure system and a wanting system that turned out to be different brain regions, with different neurotransmitters, in different sizes — the wanting large and robust, the liking fragile and small. A prescription drug for a movement disorder that turns 17% of its patients on agonists into gamblers and shoppers, validating the architecture by producing it. A priestess of Mantinea who told Socrates that Eros was the child of Resource and Poverty. A bishop in North Africa who in his fifties remembered the prayer of his twenties — give me chastity, but not yet. A French theorist at Stanford who said we do not desire from ourselves, and his undergraduate student who built the largest internal-mediation engine ever assembled. A secretary who walked up Fifth Avenue smoking on Easter Sunday 1929. Twenty-two lottery winners who turned out to be no happier than their neighbors. Two hundred and five Germans wearing beepers who recorded that the desires they could not resist were for screens and email, not for cigarettes. An algorithm published in 1997 that turned out to be the same algorithm the brain had been running for a hundred million years. A boat in a video game, on fire, spinning in a corner of an isolated lagoon. A woman in a rented villa in Italy whose lover had begun to paint and whose love had begun to die.

Every one of them — the etymology, the neurons, the drugs, the priestess, the saint, the theorist, the advertiser, the lottery winners, the beeped Germans, the algorithm, the burning boat, the woman in Italy — is the same shape. The structure of being a creature who longs for what it does not have, and which closes again behind every door it opens. The Sanskrit word for the thing is the third of four legitimate aims of human life. The Pali word is “thirst.” The Portuguese word means “longing for what may never have been.” The Latin word is etymologically about looking up at the sky and not seeing what was promised. The English word is the most-used and the least articulated of them all.

The hole at the center of every system that wants something. We have been measuring it for seventy years, theorizing it for twenty-three centuries, and naming it differently in every language. It has not moved.