by Sam Kean
In all their years of collaboration, Luria found “no distinct limits” to Shereshevsky’s memory.* The man could recite lists of thirty, fifty, seventy random words or numbers, in order, forward or backward, after hearing or reading them just once. All he needed was three seconds in between each item, to fix it in his hippocampus; after that, it was lapidary. Even more impressive, whatever he memorized stuck with him for years. In one test Luria read the opening stanzas of Dante’s Inferno in Italian, a language Shereshevsky didn’t speak. Fifteen years later, with no rehearsals in between, Shereshevsky recited the lines from memory, with all the proper accents and poetic stress. Nel mezzo del cammin di nostra vita…
You’d think Shereshevsky would have his pick of six-figure jobs, but like many so-called mnemonists, he drifted somewhat loserishly between careers, spending time as a musician, reporter, efficiency consultant, and vaudeville actor (memorizing lines was a snap). Unfit for anything else, he finally landed a job in what was essentially a neurological freak show, touring the country and regurgitating numbers and nonsense words to audiences. The gap between his obvious talents and his lowly status gnawed at Shereshevsky, but to Luria the discrepancy made sense. That’s because Luria traced both his mnemonic prowess and his employment woes to the same root cause—excessive synesthesia.
In Shereshevsky’s mind no real boundary existed between the senses. “Every sound he heard,” Luria reported, “immediately produced an experience of light and color and… taste and touch.” And unlike “normal” synesthetes, whose extra sensations are pretty vanilla (simple odors, single tones), Shereshevsky experienced full-on scenes, full mental stage productions. This became handy when memorizing items. Instead of a violet 2 or chartreuse 6, 2 became “a high-spirited woman,” 6 “a man with a swollen foot.” The number 87 became a stout woman cozying up to a fellow twirling his mustache. The vividness of each item made recalling it later trivial.
To then remember the order of such items, as in a list, Shereshevsky used a trick. He imagined walking along a road in Moscow or in his hometown (whose layout he knew by heart, needless to say) and “depositing” each image at a landmark. Each syllable of the Dante, for instance, summoned up a ballerina or goat or screaming woman, which he’d then plunk down near whatever fence, stone, or tree he happened to be passing at that moment on his mental stroll. To recall the list later, he simply retraced his route, and “picked up” the images he’d left behind. (Professional mnemonists still use this trick today.) The technique backfired only when Shereshevsky, who was rather rigid, did something foolish, like deposit images in dark alleys. In these cases he couldn’t make the image out, and he’d skip the corresponding item on the list. To an outsider this seemed like a lapse, a chink in Shereshevsky’s memory. Luria realized that this was actually less a failure of memory than of perception—Shereshevsky simply couldn’t see the image, nothing more.
Shereshevsky’s memory played other tricks as well. He could increase his pulse rate and even make himself sweat simply by remembering a time when he’d chased down a departing train. He could also (and Luria confirmed this with thermometers) raise the temperature of his right hand by remembering a time he’d held it next to a stove, while simultaneously lowering the temperature of his left hand by remembering what ice felt like. (Shereshevsky could even mentally block out pain in the dentist’s chair.) Somehow his memory could override the “this is just a recollection, it’s not actually happening” signal from the frontal and parietal lobes that should have quelled these somatic reactions.
Unfortunately, Shereshevsky couldn’t always corral his imagination or confine it to turning mnemonic tricks. When reading a book, synesthetic images would start multiplying inside his head, crowding out the text. A few words into a story, he’d be overwhelmed. Conversations took wrong turns, too. He once asked a gal in an ice cream parlor what flavors they had. The (probably innocent) tone in which she responded “Fruit ice cream,” he said, caused “whole piles of coals, of black cinders, to come bursting out of her mouth. I couldn’t bring myself to buy any.” He sounds insane, or like Hunter S. Thompson at his druggiest. If menus were printed sloppily, Shereshevsky’s meal seemed contaminated by association. He couldn’t eat mayonnaise because a certain sound (zh) in the Russian word for it nauseated him. No wonder he struggled to hold a job—simple instructions would mutate inside his imagination and stagger him.
Even the traveling mnemonist gig eventually became oppressive. After too many years of doing the show, Shereshevsky felt old lists of numbers and words haunting him, cacophonizing inside his skull, elbowing newer memories aside. To rid himself of them, he more or less resorted to voodoo, writing out the lists on paper and burning them. (No luck—the exorcism failed.) Relief came only from suppressing such memories, by training his mind to not acknowledge them. Only dumbing his memory down took the edge off.
Most people who met Shereshevsky considered him dim and timid, a bumbling Prufrock. Indeed, he considered himself pathetic, someone who’d wasted his talent in sideshows. But what else could he have done? With so many memories crowded into his skull—his memory actually stretched back to before his first birthday—his mind became what one observer called “a junk heap of impressions.” As a result he lived in a veritable haze, nearly as befuddled and helpless as H.M. or K.C. A memory that’s too good is just as broken as one that’s no good at all.
To be useful, to enrich our lives, memory cannot simply record the world around us. It needs to filter, to discriminate. In fact, while we joke about a poor memory as a sieve, that’s actually the wrong way around. Sieves let water leak through, but they catch substantial things—they catch what we want to preserve. In the same way, a mind functions best when we let some things, like traumatic memories, go. All normal brains are sieves, and thank goodness for that.
However useful, the sieve metaphor isn’t perfect. Human memory doesn’t just filter things. Our memories actually sculpt and rework and—with surprising regularity and slyness—distort what remains behind.
Even neuroscientists, who should know better, fall prey to distortions. Otto Loewi, whose dream about frog hearts helped prove the soup theory of neurotransmission, claimed to have had the dream over Easter weekend in 1920. But the journal in which he published his results, according to its records, received his initial submission a week before Easter that year. A few killjoy historians also think that Loewi didn’t rush from his bed to the lab at 3 a.m., but instead merely wrote out the details of the experiment, step by step, then resumed snoozing. Perhaps Loewi—who loved telling tales—let the demands of narrative drama mold his memory. Similarly, William Sharpe, who harvested the glands of the giant while the family stewed in the front parlor, couldn’t have done so (as he claimed) on New Year’s Day, since the giant died in mid-January. Furthermore, a colleague of Sharpe’s later claimed to have accompanied him on his clandestine errand—and also claimed that they picked through the giant’s innards not right before the funeral but the night before, around 2 a.m. Both men cannot be correct.
Why does this happen? Why do memories get twisted like metal girders in a fire and harden into the wrong shape? Neuroscientists disagree on the answer. But one theory gaining momentum says that the very act of remembering something—which you’d think would solidify the details—is what allows mistakes to infiltrate.
When capturing a memory, neurons jury-rig a connection for the short term. They then solder those connections together with special proteins, a process called consolidation. But the brain may use those proteins for more than just capturing memories; the proteins may help retrieve and replay memories, too. Consider: If you play a beep, then shock a mouse, it sure as hell remembers this. Play the tone again, and it freezes in terror, anticipating another shock. Scientists have found, however, that they can make the mouse forget that terror. They do so by injecting a drug into the mouse’s brain just before the second beep, a drug that suppresses the memory-capturing proteins. Shockingly, the next time the tone pla
ys, the mouse keeps on doing mousey things. Without those proteins the memory apparently unravels, and the mouse never fears the beep again. This implies that our brains, when recalling a memory, probably don’t just replay a pristine “master copy” each time. Instead, they might have to re-create and re-record the memory each time through. And if that recording gets disrupted, as it did in the mouse, the memory vanishes. This theory, called reconsolidation, argues that there’s little inherent difference between recording first mnemonic impressions and recalling them later.
Now, mice aren’t little humans: humans have richer, fuller memories, and our memories work differently. But not that differently, especially on a molecular level. And if reconsolidation happens in humans—and there’s evidence it does—then having to rerecord a memory each time through probably makes it labile and therefore corruptible. To be sure, we humans don’t often forget events completely, like the mice did. But we do garble details,* especially personal details, all the time. As a troubling corollary, the memories that most define us—our tenderest moments, our traumas—could be most prone to distortion, since we reminisce about them most often.
So why do distortions creep in at all? Because we’re human. Subsequent knowledge can always taint a memory: you can never remember your first date quite as fondly if that son of a bitch cheated on you later. So you retroactively retouch things and convince yourself that he mistreated you from the start. We also don’t store memories the way computer hardware does, with each datum in a well-defined location. Human memories live in overlapping neuron circuits that can bleed together over time. (Some observers have compared this to Wikipedia editing, with each neuron able to tweak the master copy.) Perhaps most important, we feel the need to save face or goose our reputations, either by gliding over inconvenient facts or misrepresenting them. Indeed, some scientists argue that the unconscious mind confabulates—makes up plausible stories to mask our true motivations—far more often than we care to admit. Unlike victims of Korsakoff’s syndrome, normal folk don’t confabulate because of memory gaps. But we do tint what we recall and suppress what’s convenient to suppress—until we “remember” what we want to, and can believe that a life-changing dream really did occur on Easter. Memories are memoirs, not autobiographies. And the memories we cherish most may make honest liars of us all.
CHAPTER ELEVEN
Left, Right, and Center
The largest structures in the brain are the left and right hemispheres. Human brains have striking left/right differences, especially with regard to language, the trait that best defines us as human beings.
The man’s name and his reasons for shooting himself—insanity? anguish? ennui?—are lost to history. But in early 1861 a Frenchman near Paris dug the business end of a pistol into his forehead and pulled the trigger. He missed. Not completely: his frontal skull bone was shattered and flipped upward like a fin. But his brain escaped unscathed. The man’s doctor could in fact see the brain pulsating through the open wound—and couldn’t resist reaching for a metal spatula.
Unsure whether the fellow would pass out, scream, or perhaps convulse and die, the doctor pressed the spatula down gently at various points and asked him how he felt. Although no one recorded the answer, you can imagine what the man had on his mind, so to speak. “J’ai mal à la tête, docteur. C’est—” Nothing had happened so far, but when the doctor pressed one particular spot, near the back of the frontal lobe, the man’s words were snapped in two: he suddenly couldn’t speak. The moment the doctor lifted the spatula, the man started up again. “Sacre bleu, doct—” The doctor pressed again, and again strangled his words. This happened over and over—each press left him sputtering, mute. The examination ended shortly thereafter, and sadly, the patient died within weeks.
A scientist named Simon Auburtin read an account of this case at a meeting of the Société d’Anthropologie in Paris on April 4, 1861. His reasons for doing so were not wholly pure. He wanted to promote the spatula-wielding doctor, a chum of his, and the case moreover supported Auburtin’s pet neuroscientific theory: localization, the idea that a different region in the brain controlled each mental function. Auburtin was especially fascinated with the localization of language, an obsession he shared with his father-in-law. (The father-in-law had been cataloguing brain lesions since the 1830s, and in 1848 had bet all comers 500 francs that no one could find a widespread lesion in the frontal lobes without attendant loss of speech.) Auburtin seized on the spatula case as the best proof yet of a “language spot” inside the brain.
Believing in localization did not put Auburtin in the majority among his colleagues, most of whom disdained localization, dismissing it as phrenology 2.0. The original phrenology movement had died in ridicule decades before, and Auburtin himself conceded that phrenologists had gone overboard in tracing things like atheism or a “carnivorous instinct” to specific head bumps; he wanted to salvage only the general principle of brain specialization. But no matter how carefully Auburtin couched his ideas, they retained the stink of quackery. It didn’t help that localization violated many scientists’ metaphysical beliefs about the brain and soul being indivisible into smaller units. As you can imagine, this wasn’t the kind of argument you could settle in an hour, and the meeting that day in April deteriorated into squabbling.
In the audience that afternoon, taking notes for the house newsletter, sat the society’s thirty-seven-year-old secretary, Paul Broca. The son of an army surgeon, Broca had come to Paris a dozen years before. At first he’d whiled away his days writing and painting; he later found a teaching job but detested it, and was so hard up that he considered striking out for America. By his late twenties he’d righted himself and found work as an anatomist and surgeon. But with every passing year he devoted more and more time to his life’s passion—skulls, which he amassed a huge collection of. More generally, Broca loved anthropology, and he’d cofounded the Société d’Anthropologie in 1859. He’d envisioned free-ranging discussions about human origins and primitive societies (and skulls), not quibbles about brain localization. Indeed, the topic held little interest for him—at least until he met Tan.
Tan’s real name was Leborgne. An epileptic since childhood, Leborgne had earned his living making hat lasts, the wooden molds around which milliners sculpted their chapeaus. But years of epileptic damage eroded his ability to speak, and by age thirty-one all he could say, in response to any question, was “Tan tan.” That soon became his nickname, and in 1840, unfit for anything else, Tan was committed to the Bicêtre, a half hospital, half nursing home, outside Paris. He did not respond well to this confinement. Perhaps the frustration of being inarticulate overwhelmed him, or perhaps, as with H.M., other patients tormented him. Regardless, Tan turned into a real prick after being committed. Other Bicêtre patients found him egotistical, mean, and vindictive; some accused him of thievery. The strange thing was that, when pushed too far, Tan could say something besides “Tan tan.” He’d scream Sacré nom de Dieu! in their faces, scandalizing everyone within earshot. But Tan couldn’t swear voluntarily, only in the throes of rage.
However vicious, Tan didn’t deserve what happened next. In 1850 he lost all feeling in his right arm; four years later his right leg became paralyzed, and he spent the next seven years confined to bed. In those days bedsores frequently turned lethal, and because Tan never soiled his sheets, nurses rarely changed his linens or rotated him. He also had no feeling on his right side, so by the time someone noticed the gangrene, it had chewed up his right leg from heel to buttock. He needed an amputation, and on April 12, 1861, his doctors presented him to a newly hired surgeon at Bicêtre, Paul Broca.
Broca started by taking Tan’s clinical history. Your name, monsieur? “Tan.” Occupation? “Tan tan.” The nature of your troubles? “Tan—tan!” Each “tan” came out pure and sweet and dulcet—Tan’s voice still sounded nice—but the absurdist dialogue meant nothing to Broca. Thankfully, Tan had become a master mime and could communicate through hand signals. For inst
ance, when Broca asked how long he’d been in Bicêtre, Tan flashed jazz hands with the fingers of his left hand four times, then his index finger once—twenty-one years, the right answer. To check whether this was a lucky guess, Broca asked the same question the next day. Just to be sure, Broca asked a third time, the following day. At this, Tan realized he was being tested and screamed, “Sacré nom de Dieu!” (When reporting this curse in his case report, Broca used euphemistic dashes.) Broca determined from these interviews that Tan, despite losing the ability to speak, could still understand language.
Per his duty, Broca amputated Tan’s leg. But the gangrene had weakened Tan too much, and he died the morning of April 17. Within twenty-four hours Broca, still mulling over the recent Société debate about the “language spot,” opened Tan’s skull.
Inside Broca found a mess. The left hemisphere looked deflated, and it almost disintegrated when touched. The frontal lobe especially looked nasty: it contained a rotten cavity “the size of an egg,” with yellow ichor pooled inside. Despite the mess, Broca’s trained eye noticed a crucial detail: that the putrefaction, while widespread, seemed to get worse the closer he got to a central point. And the bull’s-eye of this putridity sat near the back of the frontal lobe—exactly where the doctor in the case report had pressed the spatula. Broca deduced that this was the original lesion. And because Tan’s original symptom had been loss of speech, Broca concluded that this area must be a language node. In deciding this, Broca effectively threw his chapeau in with Auburtin and the neophrenologists, a dicey career move. Even more risky, Broca decided to present Tan’s brain to his beloved Société at its next meeting—April 18, that very afternoon.
