CHAPTER 10
Recovered Memories
To the boy Henry Adams, summer was drunken. Among senses, smell was the strongest—smell of hot pine-woods and sweet-fern in the scorching summer noon; of newmown hay; of ploughed earth; of box hedges; of peaches, lilacs, syringas; of stables, barns, cow-yards; of salt water and low tide on the marshes; nothing came amiss.
—HENRY ADAMS, The Education of Henry Adams:
An Autobiography (1918)
WHO HAS NOT ENCOUNTERED A LONG-FORGOTTEN odor that brings to mind suddenly, and with great clarity, a moment from the past? It leaves one marveling at the potency—and persistence—of smell memory. It’s an experience people are eager to share with me. A compilation of their stories would make a great autobiography of the nation’s collective nose. The American essayist Ellen Burns Sherman had a similar idea: “Were they all collected in a volume, what a golden treasury of poetry and romance would be the thousand records, grave, sweet and tender, which are evoked from every one’s past by the swift coupling line of olfactory association.”
Conventional wisdom credits the French novelist Marcel Proust with the first literary description of the link between smell and memory. His well-known account appears in the opening pages of his multivolume novel Remembrance of Things Past (1913), when the scent of a madeleine dipped in tea awakens childhood memories for the narrator, Marcel. A madeleine is a scallop-shaped sponge cookie—a bite-sized Hostess Twinkie without the filling, and without much flavor. That Proust constructed a 3,000-page story around it is, by itself, one measure of his literary genius.
The madeleine episode has become a cultural touchstone for the smell-memory experience. The poet Diane Ackerman calls him “that voluptuary of smell” and a “great blazer of scent trails through the wilderness of luxury and memory.” The psychologist Rachel Herz claims, “Proust may have been prescient in noting the relationship between olfaction and the phenomenological experience of reliving emotions of the past.” The science essayist Jonah Lehrer believes Proust revealed “basic truths” about memory, specifically that it “has a unique relationship” with the sense of smell. Lehrer credits the novelist with arriving at these truths before scientists did; in fact, he says “Proust was a neuroscientist.”
Psychologists have made Proust their mascot for smell memory. Psychology journals are full of brand-conscious titles like “Proust nose best: Odors are better cues of autobiographical memory” and “Odors and the remembrance of things past.” One has to admire how thoroughly Proust cornered this market—no other novelist has a branch of science named after him. Skepticism being one of the chief values of science, this sort of cheerleading makes one wonder whether Proust’s insights justify the hero worship. Was he really the first writer to note a link between smell and memory? Did he really foreshadow modern neuroscience? To find the answers, we need to look more closely at Proust’s original account.
THE ICONIC MADELEINE passage was published in 1913 in Swann’s Way, the first installment of Remembrance of Things Past. A grown Marcel is served tea and a madeleine by his mother. When he lifts a spoonful of tea and cookie to his lips, he shudders and feels an “all-powerful joy”: “An exquisite pleasure had invaded my senses, something isolated, detached, with no suggestion of its origin.” Marcel is overwhelmed by a nonspecific sense of familiarity. The smell and taste of the madeleine have something to do with it, but are not enough to evoke a specific memory. Marcel struggles to pinpoint the source of his déjà-smell. He tastes the madeleine again, plugs his ears, and tries to relive the initial experience. Finally, after two pages of strenuous effort, it comes back to him. When he was a child, his aunt Léonie would give him, on Sunday mornings, a piece of madeleine dipped in her tea.
Proust’s struggle with the soggy madeleine is distinctly not the way most people experience odor-evoked memory. For most of us, these recollections spring to mind easily. We experience Sherman’s “swift coupling line of olfactory association,” not a prolonged, constipated mental effort. The smell scholar Dan McKenzie captures the feeling of effortlessness: “This strange revival of bygone days by olfaction is…automatic. It is most clearly and completely to be realised when the inciting odour comes upon us unawares, and then as in a dream the whole of the long-forgotten incident is displayed, even although it may have been an incident in which the odour itself was not specially obtrusive.”
Here is another remarkable thing about the madeleine episode: it is utterly devoid of sensory description. Across four pages of text, Proust, that “voluptuary of smell,” provides not a single adjective of smell or taste, not a word about the flavor of the cookie or tea. This is hard to square with his reputation as the sensual bard of scent. Outside of psychology, in fact, the experts are more impressed with his visual imagery. The literary scholar Roger Shattuck, for example, thinks that Proust’s dominant mode of description is visual. Shattuck took a close look at the eruptions of involuntary memory that Proust called reminiscences or resurrections (moments bienheureux). Of eleven examples in the entire novel, only two are triggered by smell, the madeleine incident being one of them.
Victor Graham is another scholar who finds that Proust’s sensory imagery is largely visual. Graham indexed all 4,578 sensory impressions in the novel and found that 62 percent were visual. Smell and taste together accounted for less than 1 percent. This seems shockingly low, but it is on a par with other writers. In 1898 an obsessive psychologist named Mary Grace Caldwell tabulated every sensory adjective in the poetry of Shelley and Keats. She found that visual descriptors predominated: 79.9 percent for Shelley, 73.7 percent for Keats. Smell barely registered: 1.8 percent for Shelley and 2.7 percent for Keats.
Despite his reputation, Diane Ackerman’s “great blazer of scent trails” was no more nasal than the next guy; nor did he write about smells very well. As Graham pointed out, Proust liked involuntary memories because they called forth “a flood of visual images” and emotions, but the flood contained very little aroma. Proust’s trademark as a writer was to observe the recovery of a memory in excruciating detail, though after 3,000 pages it’s not clear whether Marcel even liked the taste of madeleines. He was more interested in the process of introspection than in the smells it dredged up.
If Proust’s reputation for psychological accuracy is questionable, what about the common assumption that he was the first author to recognize a powerful link between scent and memory? The record is clear, and it does not favor Proust. In American literature the memory-evoking power of smell was a commonplace observation long before Swann’s Way. Sixty-nine years earlier, for example, Edgar Allan Poe wrote, “I believe that odors have an altogether idiosyncratic force, in affecting us through association; a force differing essentially from that of objects addressing the touch, the taste, the sight, or the hearing.”
In 1851 Nathaniel Hawthorne expressed the same idea in The House of the Seven Gables: “‘Ah!—let me see!—let me hold it!’ cried the guest, eagerly seizing the flower, which by the spell peculiar to remembered odors, brought innumerable associations along with the fragrance it exhaled.”
In 1858 Oliver Wendell Holmes called attention to odor memory in his collection of essays The Autocrat of the Breakfast Table: “Memory, imagination, old sentiments and associations, are more readily reached through the sense of SMELL than by almost any other channel.” Holmes illustrated his observation with an example from his own life. It’s a sensory rhapsody of childhood in Cambridge, Massachusetts, sometime before 1825:
Ah me! what strains and strophes of unwritten verse pulsate through my soul when I open a certain closet in the ancient house where I was born! On its shelves used to lie bundles of sweet-marjoram and pennyroyal and lavender and mint and catnip; there apples were stored until their seeds should grow black, which happy period there were sharp little milk-teeth always ready to anticipate; there peaches lay in the dark, thinking of the sunshine they had lost, until, like the hearts of saints that dream of heaven in their sorrow, they grew fragrant as the breath of angels. The odorous echo of a score of dead summers lingers yet in those dim recesses.
Holmes was a practicing physician as well as a writer. From his medical training he was well aware of the neuroanatomical basis of odor perception, and he had the Autocrat himself discuss it:
There may be a physical reason for the strange connection between the sense of smell and the mind. The olfactory nerve—so my friend, the Professor, tells me—is the only one directly connected with the hemispheres of the brain, the parts in which, as we have every reason to believe, the intellectual processes are performed. To speak more truly, the olfactory “nerve” is not a nerve at all, he says, but a part of the brain, in intimate connection with its anterior lobes.
The Professor contrasts this with the wiring of the gustatory system to explain why smell has a powerful link to memory but taste does not. Holmes’s understanding of brain function is correct and modern—and it was written fifty-five years before Swann’s Way.
While Proust was working on his novel, other writers were exploring the smell-memory connection. In 1903 the American physician Louise Fiske Bryson wrote, in Harper’s Bazaar, “An odor, a perfume, will serve to recall bright scenes of other days with a vividness that is almost a miracle.” In 1908 The Spectator published the essay “Scent and Memory,” which used the image of a magic-carpet ride to describe how a sudden scent makes “miles of distance and decades of years vanish.” Five years later Proust likened smell memory to being magically transported by a genie from the Arabian Nights.
Ellen Burns Sherman’s thoroughly psychological account of odor memory was published in 1910, three years before Swann’s Way. She described how an emotional moment woven into a man’s memory along with the scent of his lover’s perfume is brought to mind decades later when he catches “an infinitesimal whiff of the fragrance.” Sherman says the former scene appears instantaneously, as if with “the turn of an electrical switch.” In 1913 the American popular science writer Ellwood Hendrick, writing in The Atlantic Monthly, said, “These flashes of memory aided by smell are wonderful. Through smell we achieve a sense of the past.”
Clearly, the subject of scent and recovered memory was very much in the air during the first years of the twentieth century. Proust shared this fascination and gave it his characteristic introspective literary treatment. For anyone not wearing Proust goggles, however, he was obviously not the first author to anticipate the discoveries of modern neuroscience.
HOW SECURE IS Proust’s reputation as an olfactory innovator, if all these Yankees were saying the same thing years earlier? Perhaps he was the first French author to capture the phenomenon? Ah, mais non. The French author Louis-François Ramond de Carbonnières (1755–1827) was well known in Proust’s day. In his most famous work, Travels in the Pyrenees, he described his descent from a mountaintop glacier on the border between France and Spain. He became intoxicated with the rustic smells of newly mown hay and flowering linden trees. As night fell, he tried to account for “the sweet and voluptuous sensation” that came upon him with such involuntary insistence. “There is something mysterious in odors which powerfully awaken the remembrance of the past…. The odor of a violetrestores to the soul enjoyments of many springtimes.” This has a Proustian ring to it, and for good reason. As the historian and critic Charles Rosen points out, “The coincidence is not fortuitous: Proust knew this page of Ramond.” It was anthologized in French high school textbooks until very late in the nineteenth century.
Contemporary French psychology is another possible source of Proustian insight. Introspection was the research technique of choice—studies were done with one or two subjects trained to report their mental experience in precise detail. This emphasis on self-observed mental processing, of narrating one’s inward gaze, is similar to Proust’s “modernist” literary style. Théodule Ribot was the founder of modern scientific psychology in France; his 1896 book on the psychology of the emotions included a chapter on olfactory memory, which had been published earlier in the widely read Revue Philosophique. Ribot discussed such “Proustian” matters as odor memory, mental imagery for smell and taste, smell dreams, and smell hallucinations. The Revue was read not only by scientists but by the educated public, and Proust, who devoured periodicals, likely knew of it.
Between 1901 and 1903 the Revue published several articles on emotional memory. One, by a twenty-one-year-old French psychologist named Henri Piéron, contained this observation: “Sometimes, when passing through a certain place, while in a certain physical or mental state, I perceive a scent that, by itself, cannot be expressed or determined, that does not fit into the classification of odors; a composite, mixed scent that suddenly and violently plunges me in an indefinable, completely inexplicable but clearly felt and recognized emotional state.” This sounds a lot like Proust’s version of smell memory—all that’s missing is the madeleine. (Piéron went on to coauthor a textbook and become un grand frommage in French psychology.)
Roger Shattuck identifies yet another French source of Proust’s inspiration. In 1896 the philosopher Henri Bergson published Matter and Memory, a treatise on psychology that gained wide public attention. The nature of memory was at the core of Bergson’s psychology, and he stressed in particular “pure or spontaneous memory,” i.e., personal memories that survive in the unconscious for a long time before being recovered. The similarity to Proust’s involuntary memory was obvious enough that Proust was asked about it in an interview in 1913. He denied being influenced by Bergson, a denial that Shattuck says “can only be termed ingenuous.”
Marc Weiner, a professor of Germanic Studies at Indiana University, offers the sinister speculation that Proust lifted the tea-and-madeleine idea from Richard Wagner. When the composer was exiled from Germany for his political activities, he was unable to find any authentic zwieback biscuits. This led to a severe creative blockage while he was working on Tristan und Isolde. One day he received a shipment of real zwieback from Mathilde Wesendonk, his muse and platonic lover. In a letter, Wagner tells her (tongue-in-cheek) of the miraculous effects of her care package; how, when dipped in milk, the zwieback cured his writer’s block and inspired him to move ahead with the opera. The Wagner-Wesendonk letters were widely read at the turn of the century; a French edition was published in 1905, eight years before Swann’s Way. Weiner mischievously suggests that Proust’s madeleine-soaking was inspired by Wagner’s zwieback-dunking.
The Proust Boosters
Though Proust’s notion of smell memory isn’t very original, that hasn’t stopped psychologists from adopting it with enthusiasm. The first researcher to charge forth under the banner of the soggy madeleine was Brown University’s Trygg Engen. In a 1973 paper in the Journal of Experimental Psychology, he said, “The Proustian view is that odors are not forgotten to the same extent as are other perceptual events. Is there any factual validity for this claim of the artist?” Engen reported that the ability to recognize a set of memorized odors, though not high to begin with, did not drop off much over the course of several weeks. He concluded, “The Proustian insight is validated!” (His exclamation point, not mine.)
Engen’s claim that odor memory doesn’t decay was newsworthy. Mainstream memory theory in the 1970s was based almost exclusively on tests using words or pictures; memory for these stimuli faded according to well-known timetables. Yet from the beginning, smell psychologists assumed that odor memory was unique, a view steeped in conventional wisdom and garnished with anecdotes. Reviewing this period, Judith Annett notes that “negative experimental results were often taken to support the ‘Proustian’ position.” The Proustian consensus that emerged in the 1970s—that odor memory decayed slowly if at all, and was unchanged by later experience—turns out to be wrong on both counts.
Engen’s notion of indelible olfactory memory began to unravel in the 1980s. Heidi Walk and Elizabeth Johns, of Queen’s University in Ontario, observed classic interference effects—smelling a second odor soon after the first makes the first one harder to remember. Others found that rates of forgetting were the same for odors as for sights and sounds. Odor memory appeared “to be governed by the same principles as remembering stimuli in other modalities.” Such principles include interference effects and so-called rehearsal effects (an improvement in memory brought about by verbally describing the to-be-remembered odor). Most subsequent research, as the psychologist Theresa White has pointed out, shows that olfactory memory obeys the same rules as memory in the other senses: it erodes with time and is muddied by subsequent experience. The purity and infallibility of smell memory—an insight central to Proust’s literary conceit—doesn’t hold up to scientific scrutiny.
HAVING ROLLED SNAKE-EYES on their first Proustian bet, psychologists pushed their chips onto another. They proposed that personal memories elicited by odor were older and more emotion-laden than those sparked by words or pictures. The new experimental strategy was to give someone a smell, ask him to come up with a personal memory about it, and then rate that memory for age and strength of feeling.
Chief among the second generation of Proust Boosters was Rachel Herz, another Brown University psychologist who in one study asserted that she had produced “the first unequivocal demonstration that naturalistic memories evoked by odors are more emotional than memories evoked by other cues.” Her bold claim deserves a close look. Herz asked people to recall a personal memory after she gave them an odor or a picture. People then rated their memories for emotionality. Picture-prompted memories had lower emotionality scores than odor-prompted ones, giving rise to Herz’s claim. What she glosses over is the fact that both types of memory scored below the midpoint of the rating scale. In other words, visual memory and odor memory were both on the unemotional side of the scale. The odor-cued memories were simply less unemotional.
The Swedish psychologists Johan Willander and Maria Larsson have failed to confirm Herz’s results. They cued autobiographical memories with odors, words, and pictures, and found that picture-evoked memories were the most emotional and odor-evoked ones were the least emotional. Willander and Larsson write that “we did not find support for the notion that olfactory-evoked memory representations should be more emotional than memories evoked by other sensory cues.” It now looks as though the modified Proustian hypothesis—that odor memory, while not indelible, is more emotional—doesn’t hold up too well either.
By 2000, the third generation of Proust Boosters arrived and wasted little time before turning on their predecessors. The British psychologists Simon Chu and John Downes criticized previous studies for being insufficiently Proustian. (They pointed out, for example, that the memories examined in some experiments were not truly autobiographical.) Chu and Downes contrasted those failed attempts with their own research agenda, which, in their modest view, captured the true spirit of Proust. Their goal was nothing less than “translating the essence of Proust’s anecdotal literary descriptions into testable scientific hypotheses using the language of contemporary cognitive psychology.” (This is a patently ridiculous thing for scientists to do. How can a work of fiction, no matter how well written, become the truth standard for scientific research? What’s next? Will sex researchers lift hypotheses from Danielle Steel? Will Stephen King inspire psychiatric theories of fear?)
From out of left field came a quick challenge to Chu and Downes. J. Stephan Jellinek is a German psychologist who has worked as a perfumer and fragrance marketer. Not being an academic, he had the temerity to ask whether lab studies that relied on contrived and twice-prompted memories could capture the Proustian experience in any meaningful way. From a close reading of the madeleine episode, he extracts nine specific and testable characteristics of that experience. (Most have to do with the difficulty in identifying the emotion, tying it to an odor, and connecting the odor to an event in the past.) According to Jellinek, the experiments of Chu and Downes address only three of the key characteristics. Does measuring emotional response on a seven-point rating scale, he asks, truly capture the ecstatic experience described by Proust?
Determined to prove that odor memory is distinctive in some way, the latest Booster studies now claim that odors evoke older autobiographical memories than do words or pictures. This is an intriguing but ultimately trivial proposition. Whether this claim—the latest in a series of special pleadings—holds up is almost beside the point. Whether a lab experiment has captured the essence of Proust is certainly beside the point. The bigger question is why investigators decline to observe the natural history of smell for themselves, and prefer to base their research on a fictional episode. Three generations of psychologists have done so, and in each case they got lost in the woods. In the 1970s and 1980s the Proust Boosters grossly overestimated the permanence of odor memory. In the 1990s they overstated its emotional content. In the new century they overplayed how well lab studies could mimic an episode of fiction. Perhaps it’s time for them to set aside the soggy Twinkie.
MEANWHILE, OUT IN the real world, a lot of people think that odor memory is special. A Norwegian survey recently compared popular beliefs with scientific findings regarding memory. Among the general population, 36 percent believed—incorrectly—that smells were remembered better than sights or sounds. This may reflect the fact that there is something unsatisfying about the current scientific view. If odor memory is like other forms of memory, why does it feel so magical when a sniff triggers a twinge of remembrance? A lot of it has to do with surprise. You weren’t trying to remember the paints, oils, and solvents in Grandpa’s workshop—the memory popped up, unasked for, when you walked through a random odor plume. Even more surprising: you never made a deliberate effort to memorize those smells when you were seven years old. If you had, the recollection would be no surprise. In grade school you memorized the state capitals; to recall one years later doesn’t feel magical. Because odor memories accumulate automatically, outside of awareness, they cover their own tracks. We don’t remember remembering them. The sense of wonder that comes with the experience is, like all magic, an illusion based on misdirection. Like a nightclub mentalist, the mind presents us with a memory it picked from our pocket when we weren’t looking.
Henry Adams: The American Alternative
Psychology’s preoccupation with Proust has led to a narrow emphasis on involuntary memory and a neglect of the far more common features of the mental smellscape. These include how and why we willingly commit some smells to memory and not others; how and how well we retrieve them; and how fully we are able to reexperience them. These questions are a promising starting point for a fresh exploration of olfactory memory.
If Marcel Proust is the poster boy for private, involuntary odor memory, this new alternative view will need its own mascot. I propose the American author Henry Adams, who conveyed in one sentence the actual sensations of a childhood smellscape. In his autobiography, written in the third person, he gave us a litany of scents from a boyhood in the days before the Civil War. As we return with him and stand beside the barefoot kid of summer, we feel his love of the outdoors: not for him the scent of inky copybooks or Mama’s perfume, lavender sachets in the linen closet or bread in the oven.
Henry Adams gives us a small sample of a true olfactory memoir—it puts you behind another person’s nose in another time and place. In his honor, I call it Adams-style odor memory. To my way of thinking, Adams-style memory beats Proustian memory because it deals with smells that are deliberately sniffed and voluntarily recalled. These are not the buried land mines of Proustian memory; Henry Adams describes a smellscape that was familiar to his entire generation, and his memory of it is open to the public. Proustian memory inhabits a private, interior place, and is open by invitation only. For Proust, smell was a tool, a reflex hammer he used to probe his own mind. For the young Henry Adams, smell was the whole world; for the old Henry Adams, it was an open gateway to the past. Breathe deep: it’s summer, the sun is hot, and the tide is low.
Adams-style odor memory is popular with American writers. A fine example is found in the opening lines of Lake Wobegon Days, where Garrison Keillor conjures up the fictional town of Lake Wobegon, Minnesota:
Along the ragged dirt path between the asphalt and the grass, a child slowly walks to Ralph’s Grocery, kicking an asphalt chunk ahead of him. It is a chunk that after four blocks he is now mesmerized by, to which he is completely dedicated. At Bunsen Motors, the sidewalk begins. A breeze off the lake brings a sweet air of mud and rotting wood, a slight fishy smell, and picks up the sweetness of old grease, a sharp whiff of gasoline, fresh tires, spring dust, and, from across the street, the faint essence of tuna hotdish at the Chatterbox Cafe.
You don’t have to be a Norwegian bachelor farmer to appreciate this. Anybody can inhale the scene and experience Lake Wobegon.
Adams-style memory has a big scope: it’s about extended episodes, not single events, entire smellscapes rather than isolated odors. Adams-style memory edits an entire season down to an aromatic highlight reel that can replayed at will. Dozens of Saturday afternoons with Grandpa at his workbench are distilled into a few key molecules.
By preserving familiar scenes, Henry Adams left us a time capsule of a lifestyle that has nearly vanished. For most of our history, most Americans lived and worked on farms; agriculture was our common smellscape. Haydn Pearson was born in 1901 and grew up on a small family farm in Hancock, New Hampshire. In a memoir, he recalls the ambience: “When I was a boy, one of my favorite spots was the livery stable. When I walked into Woodward’s Livery behind the Forest House Hotel, I was met by a pungent heady fragrance compounded of hay, leather, grain, harnesses, stained and splintered floor planks, and manure.” The interior of the livery office had its own character, “the fragrance of felt leggings, rubber arctics, sheep-lined coats, and the sawdust box for tobacco juice blended very pleasantly with the over-all aroma of the establishment.”
His family stored root vegetables and preserved foods in the farmhouse cellar, which acquired its own atmosphere: “a heavy damp pungent smell compounded of moist soil, potatoes, apples, carrots, turnips, salt pork, cold crackling brine, and the old floor boards. Probably there were some rotten potatoes and possibly a decayed cabbage or two, and if there is any farm-cellar fragrance equal to the combination of decayed potatoes and decomposed cabbages, I have yet to smell it.”
For Ben Logan, born in 1920 and raised on a small farm near the Kickapoo River in southwestern Wisconsin, haying time was aromatic: “A time like that comes back now sharp and real with all its smells of dust, horse sweat, man sweat, Lyle’s oozing pipe. There is the dry whirring of grasshoppers, steel wagon wheels ringing on the hard ground, the creak of the hay rope. There is the tepid smell of water as we drink from a bucket that has a taste of leftover lemonade. Above all is the sweet smell of curing hay.”
Proustian memory is involuntary; we have no control over its recording or its recall. Because it is recoverable on demand, Adams-style odor memory is a more useful storage medium—it embodies our common past and gives us a way to preserve it. Some people improvise their own olfactory scrapbook. An attorney who was in his thirties at the time once described his method for inducing scent-fueled visions of the past:
I grew up on the Nevada desert in a small mining town. Since my seventeenth year my residence has been in California in the San Francisco bay area but I never have and never will learn to be happy in the fog and rain and dampness. I have a perpetual nostalgia for the sun, warmth, clear, clean air, the peculiar lemon desert fragrances and the great panoramic vistas and strong colors. I have spent part of several summers in the Tahoe district and each time have brought home a good bunch of sage brush which I keep in a receptacle and not infrequently smell. When I do, visual and emotional sensations arise within me in considerable clarity of the desert scene. A slight sniff doubles and redoubles that tranquil nostalgia.
The scientific study of smell memory is currently in flux. After a long and fruitless detour spent quantifying a literary fiction, the field is abandoning the idea that smell is unique among the senses. Just as the larger field of memory research has retreated from the notion of indelible flashbulb memory and questioned the veracity of eyewitness testimony, smell experts are recognizing that memory for odor is like memory for anything else—subject to fading, distortion, and misinterpretation. With this realization, we give up some long-held ideas, but throw open the windows for a breath of fresh air.