Surfaces and Essences

by Douglas Hofstadter

  If conceptual encoding did not take place, it would be impossible to retrieve events stored in one’s memory. Just think how indispensable it is for users of Web sites on which photos or videos are shared that linguistic labels (“tags”) are attached to each item. There are many additional arguments for the necessity of encoding in the act of committing experiences to memory, based on such phenomena as selective forgetting, partial or distorted recovery of memories, and the reconstruction of memories.

  No purely image-based search process, no matter how sophisticated, would be able to “see” the connection between Dick at Karnak and Danny at the Grand Canyon, or between Doug’s disappointment about subscripts and Monica’s disappointment about button #2, because such events’ connections are not visual. The moral is that we do not store in our memory a collection of “objective” events through which we run, seeking perceptual resemblances, whenever a new event happens to us; rather, events that befall us get encoded — that is, perceived, distilled, and stored — in terms of prior concepts that we have acquired. Remindings are possible because certain aspects of long-ago situations were noticed and stored at that time, yielding encodings of those situations. But which aspects get paid attention to, and how are they encoded?

  Do Our Brains Instantly Pinpoint Timeless Essences?

  The activation of certain memories from one’s earlier life is not a mental game that we humans engage in merely because we find it intellectually pleasurable to connect present and past through the spotting of similarities. In general, the automatic behaviors that we engage in play a crucial role in ensuring our survival. Being reminded of a past event is not a luxury add-on that might optionally take place after we have understood a new event; rather, such a reminding is deeply implicated in the very act of understanding the new situation. This idea was stressed early on by cognitive scientist Roger Schank, one of the pioneers in the study of reminding.

  Let us restate this more concretely. Doug was reminded of his son Danny at the Grand Canyon when he watched his friend Dick at Karnak because his perception of the fresh new Karnak situation activated certain concepts that had been encoded in the course of “processing” the Grand Canyon event many years earlier. Doug knew he was in a very special sacred place when he was at Karnak. He noticed that a companion had leaned down to pay attention to something small and was momentarily ignoring the guide. His emerging sense of irony in this situation led him to “replay” the scenario of Danny at the Grand Canyon, as it existed in his memory.

  Does this mean that the category trivial side show more fascinating than the main event was necessarily created at the moment the first event was encoded? Does it mean that Doug, when he experienced the second event and encoded it, automatically reactivated the category’s founding member because the two had been encoded identically? This is the idea on which Schank’s theory of the mechanisms underlying reminding is based. However, is that the only way to explain the bubbling-up of a dormant memory?

  Does a successful reminding presume that both events were encoded at such a high level of abstraction that the two situations are simply specific cases of the abstraction? Can one be reminded of a past situation only if, already in the moment when one was experiencing it, one succeeded in putting one’s finger on such an abstract “timeless essence” that later, many years down the pike, when another situation sharing that same timeless essence comes along, one will be all prepared to activate it? If this were the case — let’s call it “instant pinpointing of timeless essences” — then Doug, when he observed Danny’s fascination with the ants and the leaves, would have instantly created, albeit unconsciously and without an explicit linguistic label, the abstract category trivial side show more fascinating than the main event. His sudden recollection, as he watched Dick at Karnak, of his son at the Grand Canyon would have been due to the creation of this category some fifteen years earlier. If one were to believe the hypothesis of “instant pinpointing of timeless essences”, then such a reminding could have taken place only if the conceptual skeleton trivial side show more fascinating than the main event (we stress that it’s not the sequence of English words that we are talking about, but the abstract idea that it denotes) had been created at the moment of encoding the first scene, and then, fifteen years later, rediscovered in another scene.

  To put it mildly, this is most implausible. The instant spotting of a shared timeless essence is not the only mechanism that could, in theory, give rise to a reminding sometime down the pike. The problem is not that it’s hard to perceive a deep and precise conceptual skeleton, for in fact we all perceive such skeletons all the time, when the categories involved are those of our day-in-day-out thinking — that is, categories with which we have some degree of expertise. The problem is that when the categories are unfamiliar to us, we cannot see nearly as deeply into situations involving them.

  We effortlessly and unconsciously perceive certain shared aspects of situations that are very different — for example, despite their enormous differences, we recognize all sorts of elephant-in-the-room situations, once-bitten-twice-shy situations, you’re-pushing-your-luck situations, can’t-see-the-forest-for-the-trees situations, if-it-ain’t-broke-don’t-fix-it situations, killing-two-birds-with-one-stone situations, a-bird-in-the-hand-is-worth-two-in-the-bush situations, you-can’t-have-your-cake-and-eat-it-too situations, you-could-hear-a-pin-drop situations, people-who-live-in-glass-houses-shouldn’t-throw-stones situations, and so forth — and the recognition of these highly abstract qualities allows us to spot commonalities in situations that have nothing at all in common on their surfaces. In a word, we are all experts at perceiving conceptual skeletons when the categories involved are ones in which… we are experts!

  And so, could it be that Doug unconsciously perceived the conceptual skeleton trivial side show more fascinating than the main event while watching Danny, and used it to encode that event? Could it be that this made him an expert with this concept? Could it be that thanks to his expertise with this concept he was able to recall that event many years later, while visiting Karnak? Well, it’s possible in theory, but it’s most unlikely.

  Spotting the essence of something that one is already an expert in is one thing, but spotting the essence of something novel and unfamiliar with is quite another thing. To put it more pithily, abstraction is one thing, but deep abstraction is quite another thing. The hypothesis that we can instantly spot “timeless essences” leads, in certain cases, to absurd conclusions. When experiencing an event, we cannot peer into the future and clairvoyantly guess exactly which highly abstract encoding to construct so as to allow that memory to be triggered by events that will take place many years down the pike.

  More concretely put, when Doug was eight years old and he asked his father what the meaning of subscripts was and his father’s answer gave rise to a great feeling of loss, the abstract conceptual skeleton that this event would wind up sharing, forty years later, with the episode of Monica and the Dustbuster could not possibly have been clear to him at that early point in his life. What, then, did the eight-year-old child encode at that time? Alas, one would need both a time machine and a mind-reading machine to give a precise answer, but there is no doubt that some abstraction was involved in the encoding. The crux of perception — even for an eight-year-old or a two-year-old — is the act of abstracting; abstraction is the principle that allows us to create new categories and to extend them throughout the course of our lives. We are forever extending our categories because the strictest form of literality (“total rote recording”, as we dubbed it above) does not allow any resemblances to be noticed, and thus excludes all thinking. In the case of eight-year-old Doug, there had to be at least enough abstraction in his perception to allow the connection with the Dustbuster scenario forty years later to be perceived, even though that scenario was, in so many ways, vastly different from it.

  First Encodings Can Go Only So Far

  In a course one of us taught, a student heard the anecdote o
f newlywed Carol signing with her maiden name and her friend Peter who “did the same thing every January”; this reminded the student of a time four years earlier, when she had changed jobs. In her new job, she was supposed to answer the phone and say, “Hello, this is Company X”, but in her first few weeks she would often say the name of her previous employer, upon which she felt very silly. Does the fact that this student hit on that memory mean that four years earlier, she had constructed the highly abstract category situations in which one is trapped by a habit that one is unable to update? It’s hard to imagine what would have driven her to create such a high abstraction; it’s unlikely that she felt that someday she would need to be an expert in the spotting of situations of that sort. One can get along quite well in life without having constructed such a narrow category. Although we can be pretty confident that this student didn’t encode her mildly embarrassing phone-answering gaffes at such a rarefied level of abstraction, her encoding nonetheless involved a fair amount of generalization, because hearing the story about Carol’s maiden-name-signing gaffe brought her own old gaffes back to her mind swiftly and effortlessly.

  Among the likely aspects of her encoding were the fact that she blurted out the wrong thing, that what inadvertently came out of her mouth was a relic of her recent past, that she wasn’t in control of what she said, that she was embarrassed by her slip, and so on. And so it seems probable that her retrieval of that four-year-old memory, triggered by the anecdote of Carol’s signing error, didn’t involve just one single and concise abstract conceptual skeleton that perfectly matched both the old and the new situations, but rather, that it depended on a number of separate, abstract, and important aspects of her old job situation, each of which had contributed to how she had perceived that situation’s essence at the time it happened. In other words, rather than one single perfectly-fitting conceptual skeleton — the magic key to memory retrieval — having been created when she made her telephonic errors, a number of smaller and independent concepts were built, and hearing the story of Carol reactivated enough of those concepts to remind her of having said, inappropriately, “Hello, this is company X”. In sum, although there is always some degree of abstraction in the act of encoding, there is not always exactly one highly abstract conceptual skeleton shared by the triggering event and the event retrieved from memory.

  For a current event to trigger the recall of a far-off event that one hasn’t thought of in many years requires strong resemblances. For the long-buried memory to be triggered by what one is currently experiencing means that each side of the connection has to “give” a little — that is, some dimensions of the way we perceive the new and the old situation have to have sufficient flexibility. Even if the encodings of the events are far from reaching the maximal level of abstraction, they will go far beyond the most literal details of the experienced situations, because such literality confines remindings to the level of the very mundane. For example, in the case of the children in the Cinque Terre who were fascinated by grasshoppers, it takes only a tiny amount of abstraction to link this situation with Danny at the Grand Canyon, since both situations involve children, insects, and famous tourist spots. One feels that the stories are so close to being carbon copies of each other that such a reminding verges on the trivial. Luckily, the human mind goes way beyond this low level of reminding in the mental leaps that it carries out among countless stored memories.

  In conclusion, is there always some amount of abstraction involved in the encoding of memories? To be sure. But does being reminded by a current event of a past event always depend on the two events exactly sharing a conceptual skeleton in one’s mind? By no means.

  The Humble List that Aspired to Become a Magnificent Category

  A local chef has just finished sautéing a fresh fish caught on a small fishing boat based in a nearby village. The side dish consists of steamed vegetables grown on a local organic farm. Meanwhile, elsewhere on the same planet, a microwave oven has just heated up a frozen dinner featuring fish grown in a factory and stuffed to the gills (literally!) with bone meal, accompanied by genetically modified vegetables that were grown in chemical fertilizer laced with pesticides.

  Which of these two dishes appeals more? It’s probably not too hard to choose, but unfortunately, the first is expensive to produce and certainly is not amenable to mass production, and in addition it requires people who are passionate about what they do. Today it’s nearly impossible to make a living in such an authentic, old-fashioned way. And so, is there any way to successfully combine that which allures but is unprofitable, on the one hand, with that which is cheaply manufactured though unappetizing, on the other? Well, a good strategy would seem to be that of mass-producing something that emanates downhome appeal — if this is not a contradiction in terms.

  Below we list some cases where marketing trickery can give rise to a false impression that tempts customers to buy without realizing what lies beneath the surface. Naïve would-be buyers, convinced that they have hit on genuine authenticity, fall for slick mass-produced articles — a successful ploy resulting from cold calculations in the business office.

  Snails labeled “escargots de Bourgogne” seldom hail from Bourgogne. Today, this classic dish, so redolent to so many of France, is mostly imported from Eastern Europe, Turkey, or China, and when one is eating a snail, the chances are very slim that it actually grew in the shell in which it is found.

  In a now-defunct chain of American bookstores, certain prominent sections used to be dedicated to “Local Authors”; this gave the impression that that very bookstore’s staff had played a role in the selection of the books found in that section. But in fact, the choice of which local authors to showcase was made far away, in the chain’s national headquarters, without any input at all from the local store.

  Ads for certain little clay figurines representing historical figures or quaint folk icons proudly proclaim, “Hand-painted”, which projects a sweet old-fashioned postcardlike image of how these statuettes came to be. But the truth is that “Hand-painted” usually means “Made in China”. The statuettes of the little Danish mermaid or of the Napoleonic soldier seem far more exotic to the person who paints them than they are to the person who buys them.

  These days, on the Venetian island of Burano, known for centuries for its intricate lacework tablecloths, blouses, scarves, doilies, and so forth, nearly all such items are in fact made in China and are exported to Italy. The appearance of authenticity is preserved, however, by the elderly women selling them, who sit in the tiny shops, wearing lace clothing and working away on lovely lace items.

  In a town along the Nile in Egypt, a child comes up to tourists and offers to sell them “antiques” that she claims she found, when in fact they are mass-produced objects made for tourists and artificially aged in sand and water. In a small shop in the same town, a young boy is making scratches on a metal tray decorated by a machine, in an attempt to give the impression that he himself made all the decorations on it, and thus that all the trays have been locally decorated.

  In a Christmas market in the main square of a small Austrian town, a “Corsican peasant” dressed in a traditional Corsican costume is selling “Corsican salamis”, but his only connection with Corsica is that he likes to vacation there. Next spring he’ll don a costume from the Auvergne and bald-facedly hawk “Auvergne cheese”.

  In Cabourg, on the coast of Normandy, a crêperie on the main street is owned by a Parisian couple. On weekends, they rise very early and leave for Cabourg a few hours before the tourists do; symmetrically, a few hours after sales are over, having cleaned up and waited for the traffic jams to clear up, they head back home to Paris.

  The recent upswing in popularity of sushi bars in Paris has coincided with a downswing in the popularity of Chinese restaurants. As a consequence, Chinese restaurateurs have opened up Japanese eateries. For unsavvy Caucasian customers, nothing tips them off that the Asian people waiting on them are as out of place as a flotilla of Greek serve
rs would be in a French restaurant in the middle of Tokyo.

  The items in this list all clearly share some quality, but that quality has no standard name. What, if anything, is the difference between them and the members of a category that does have a standard name? We would suggest that the items in this list implicitly define the boundaries of a category that is just as reasonable and intellectually appealing as any lexicalized one, such as forgeries, which is a category clearly related to this one but more general than it.

  The Humble Category that Aspired to Acquire a Label

  Our list implicitly defines a new concept whose only distinction from lexicalized categories is that it is not yet lexicalized. And so, grafting “faux” onto “authenticity”, we suggest the term “fauxthenticity”. This act of explicit labeling will help to anchor the concept in memory and will increase the likelihood of its being further extended in the future. Indeed, instances of fauxthenticity are easily found, sharing the conceptual skeleton that emerges from reading the previous examples, while also extending it in new directions.

  On envelopes one receives these days, it’s common to see one’s name and address in what looks like handwriting, when in fact it has simply been computer-printed, using an intentionally informal-looking cursive font. There are even certain fonts that have randomizing algorithms in them that allow each letter token to be slightly different from other tokens of the same type, thus giving the impression that every letterform has been uniquely penned by a human hand.

  In the automated telephone trees that one encounters whenever one calls any large business, a new trick has been added fairly recently — that of having the recorded voice insert hesitant pauses, or make slight mistakes, or even sound surprised by an afterthought, as if the “person” had just remembered something they should have said earlier. “Uhh… oh, yeah. So now, could you, um, just tell me your confirmation code one more time? Thanks a lot!”