People who study memory are familiar with the concept of a Memory Palace, a memory scheme for storing information in remembered places. The Memory Palace story is beautifully described in Joshua Foer’s book Moonwalking With Einstein (although it wasn’t clear to me if Einstein was walking on the moon or walking with Michael Jackson). For millennia, mnenonic super stars have memorized gargantuan lists by associating each list item with a location in a virtual home or palace. During learning, the mnemonicist imagines walking through the palace and placing each item in a particular location. At the time of recall, the memorizer imagines walking through the memory palace and, as if by magic, seeing each item where it had been placed. According to this notion, the ability to associate places to content is natural and fairly easy to master. So easy that Joshua Foer, who had no obvious memory potential, became a national memory champion by mastering the technique. I find the accounts impressive and extremely convincing, although I’ve failed miserably when trying to do this. Overall, the Memory Palace phenomenon suggests a strong natural link between location and memory. Some have taken this linkage to suggest that place serves as in indexing system for memory — a kind of lookup table. A point I’ll return to.
My personal experience points the other direction. Rather than a place-to-percept association, my experience suggests strong and effortless percept-to-place associations. I’ll recount two stories; you can supply your own. Last summer, while taking a long walk with my dogs, under utility lines behind my house, I listened to an album on my iphone, one I had never listened to before, and didn’t listen to again for awhile. About 6 months later I was in my car and listened to the same album. As each song played a vivid image of exactly where I had been along the walk popped into my head. The images were very clear, effortless, dramatic and beautiful. Second anecdote. I’ve been interested in the concept of information for years. This morning I was discussing my fascination with the concept, and another dog-walking memory popped into my head. A very clear evening walk, with another dog (deceased). It was along that walk, about 10 years ago, that I started to generate my notions of information. What was remarkable this morning was the clarity of where I was on the walk: the place, the weather (spring evening). My thought-walk covered about half a suburban block.
I ask reader to come up with their own stories. Among the clearest are music associations. The phenomenon of “I remember where I was when I heard that song …”. Or the smell place association, “I remember that smell from my middle-school cafeteria …”. If there is serious scientific data about this, I’d like to know.
My current notions are:
- the association between place and content (especially memory) is a two-way street
- the content->place association seems stronger than the place-> content association.
Penfield’s Hippocampal Lookup Table. Drawn on a napkin for Brenda Milner.
Place as a lookup table. The overall conclusion is that place and memory form strong two-way associations. Based on O’Keefe’s discovered that neurons in the rat hippocampus act like pointers on a map; O’Keefe and Lynn Nadel proposed that the rat hippocampus is its Cognitive Mapping System. The relation to other content and memory is speculative but strong. For example, the paper Understanding memory through Hippocampal Remapping. Briefly, the core mapping system of the hippocampus serves as the structure of a look-up table for episodic memory. There is a natural associative process between memories and location. We use this process for memory access. The process is similar to the hippocampal lookup table proposed by Penfield. Hence the relation between hippocampal damage and amnesia.
Update. PatienRB reminds me of something I wanted to add: how pleasant place-associated memories can be, at least for me. One set are the memories associated with music. For me, mostly the music of the 50s, 60s and 70s. During the ’50s I was in grade school; pop music, largely early rock, was around, especially on the car radio. Now, if I hear a Buddy Holly or Fats Domino song, I recall the car, and sometimes the route and an inkling of my 8-year-old self. Place comes first. The overall point being that evoking the memories of place elicits glimpses of the story of my life, and I have a serene sense of my continuous, autobiographical self.
Corona Radiata 
My ex girlfriend, who lives in London, came to visit me last year to see if we could still work. We spent ten days in a swanky Manhattan hotel and did nothing else than being intimate in anything and everything. This time holds precious memories. (And although we could obviously still work, the pond quickly shut down any hope for the two of us). The hotel’s complimentary shampoo had a distinct smell. I kept a vial.
Everytime I smell the shampoo I at once get a surge of memory. First, the hotel room, particularly the bathroom. Then follows a staccato of images from that whole ten day. I have not determined if there’s repeated sequence to those, of either a spatial or temporal nature (and perhaps trying to do so would spuriously contruct an artifactual one, since being introspective, the measurement may change the measured).
Concerning the two-way street, which our subjective experience makes convincing: the data structure of the memories is a factor to consider if any assymetries are to be understood. The features of what is “space” are different from those of other contents. Following thought from antiquity onwards, space is anisotropic; moving ten paces back from a place that took ten paces to get to brings you back to your starting place. Places are thus sequenced like 1, 2 and 3.
Hence a memory organized as a spatial data stucture may possibly be used as a look-up table but there may be also a bias against random acess in favor of sequential retrieval, whose delay brings a cost that memory about smells won’t have (spatial orgatization of olfactory data may exist, but i’m not aware of any scheme).
Have you read “The Art of Memory” by Frances Yates? She documents mnenonics since antiquity, at which time they were developped very extensively so that officials, particularly in the Roman world, could remember and organize the vast amounts of information needed in the bureaucracy of a 50-million souls Empire without Intel or Google. They became like Dune’s “Mentats”: humans computers (as a parallel in Dune computers are illegal).
Thanks John for this very nice blog.
PatientRB, I haven’t read Frances Yates. Thanks for the tip! I doubt memory is “random access”; It would be dysfunctional to have totally accurate memory; this would likely preclude rapid, creative association. Each memory would be accurate and specific, but isolated. The nature of association, whether it works by place, place proxies, or other factors, such as time, is fascinating and important.
I have a copy if you want to check it.
Random and sequential access mechanism may coexist. A random access could be followed by a propagation of sequential retrieval. This is present in programming languages, to one extent or another. While random access is prevalent some languages use sequential retrieval as a highly efficient way to replace looping. It’s interesting that the primary examplar, LISP, was originally favored as the language of artificial intelligence.
I wrote anisotropic. It should read isotrpic.
Pingback: There are places I remember … | Corona Radiata
I have a very spatial memory. My wife will ask, “Do you remember the pie we had at so-and-so’s house a few years ago?” and I won’t. I won’t remember so-and-so’s face, or what we discussed, or what the house looked like. But I do remember where in the house the kitchen and dining room are, and where the house is in town.
My spatial map can get mixed up, however. When I moved to Boston, I was riding the T in and popped up in Copley Square to get a cell phone for apartment hunting. For some reason, the orientation of those blocks got flipped around in my brain. To this day, whenever I walk or drive through Copley square, I have to pay extra attention and think about the direction I’m headed, or I end up turning the wrong way.
I’m spatial, too. One test I’ve done with my sisters and a few close friends who went to the same grade school is this: “try to remember the room location of your classroom and teacher for each year in school from kindergarten thru 5th grade.” Not sure what it shows, but most can sort-of do it. I can do it fairly easily, and, while doing it, picture the inside of each classroom. Sometimes playing in the mind’s attic is fun.
If you know about the “memory palace” concept, you probably also know about songlines:
http://en.wikipedia.org/wiki/Songline
Humans have been finding their way through the world with song for as long as there has been an oral tradition, and it’s likely that before that we used our hands to the same effect. Persistence hunters use their hands to signal to each other what the hunted animal appears to be up to, for example, and just being able to do this requires the ability to associate memory with hand gestures.
My personal experiences also include associations of emotional events with particular weather. Some songs are also so powerfully evocative of a particular time for me that I cannot listen to them without some significant recurrence of traumatic experience (maybe I’ve had a rough life…).
Wayfinders use stories to find their way back through the forest. This is one of the advanced techniques of wayfinding, but again, rooted in songlining. I have done this, and it’s a very good skill to have if you wander off trail a lot.
So yes, it has to go both ways, I think. But I agree that it’s probably more embedded in our skill set to create the song to help us remember the place than it is to create the place to help us remember the song. The latter is the storytelling skill. The former is the wayfinding skill. Both have been intrinsic to our way of life for a damned long time. 🙂
I’m interested in how we can use these skills to see in the other direction: the trajectory of future events.
Thanks, Leha,
I didn’t know about ‘song lines’. Fits with my experience, but adds a lot to it. The big question: why? Why this relation between songs and memory? An advantageous, but lucky connection (Steven Jay Gould’s spandrel) or an evolved trait?
John, I’m convinced it’s an evolved trait. Or at least that the capacity to remember songs and associate them with place is evolved from the practice of moving through forests while searching.
Humans and their ancestors have had to find our way in the wild for a lot longer than there has been written language, or even the concept of a crude map. The ability to associate the songs of not only other humans, but the wild world in general is absolutely intrinsic to understanding the lay of the land around us. We have used every sign we could track in persistence hunting, as well as in gathering of plant foods. Sensory awareness led to symbolic awareness, through footprints and animal sounds seen and heard, then recorded in imitations and drawings. We had to be intimately aware of subtle differences in the lay of each footprint, or the various shapes of birdsong, in order to understand what was happening around us, and this meant that, especially while hunting, we often had to fly through the landscape as fast as our feet could carry us, while simultaneously reading the signs all around us. Thus we cultivated the brain circuitry to read and to process sound and rhythm (bird calls and footfalls) in unison. Remembering what all this meant was our key to a successful hunt, as well as to finding our way, and the logical underpinning for reversing the process, using the lay of the land (or reading) to help us remember, or even to create memories. How could we teach our children to hunt without these songlines to find their way in the forest?
I’m certain that the ability to process written language–to read–was originally the ability to read the tracks of animals. That is, for me, the birth of symbolic language, and the auditory product of such track-reading is very likely the birthplace of song.
At least, that’s my take on it. 🙂
Language, music, singing, are (or seem to be) clearly linked. For example, the ability to hold a sound sequence in your head — about 3 or 5 sec of iconic memory — is (if I remember) much greater in humans than apes. Not clear what the driver, the specific evolutionary pressure, for this was. But it ends up being useful for a variety of functions. In a similar vein, symbolic representation. Is it a specific trait, with evolutionary advantage, or a free add-on that goes with having a big brain, that may have evolved for other purposes?. I’m cautious about seeing the adaptive value in useful traits.
I guess I’m kind of throwing caution to the wind. But I’m an artist, so I get to do that. 😉
Many Scientists are in your camp. I tend to be cautious.