They're doing research with mice again; though, it isn't like they've ever stopped, but in this case, they're looking at memory. Scientists are watching the hippocampus as they subject the little rodents to a series of tests (dramatic events of the sort that make a lot of little lights go off on their monitors), and are learning how the brain registers the sensory and neurological details of these events. They have discovered that each record of an event (each "memory") is a pyramid-shaped cluster of aggregated triggers (Tsien and Osan call the sub-set of neurons that register a specific characteristic of the event "neural cliques"). These pyramid collections are then transferred into long-term storage via a method of "re-experiencing" each event. But, it is these neural clusters that have fascinated the researchers because they imply that memory is not an unique record of an event, but rather an assembly of attributes.

Somewhat like saccadic masking—how we see isn't an accurate representation of external reality, but is actually an extrapolation done by our brains with the rapid-fire still images that our brains record. Memory is the same thing: attributes—binary switches—that are organized and interpreted by our brains into something approximating an accurate record. They form a duality matrix, though one that is more grounded in scientific observation than psychological organization.

Memories, then, become an assembly of binary data sets, and given an observable system, it is possible that other triggers—external ones, even—can be tripped by the presence of certain binary sequences. In other words, it may be possible to send a key sequence to an electronic lock by virtue of recalling a certain memory that exhibits a specific binary sequence. Oh, yes, your memory of eating ice cream on the porch of your grandfather's house? Yes, 00010011 is correct. Access granted.

Where 00110011 may be a different ice cream memory. One that has the additional attribute of having a dog in it. The neighborhood mutt, say, the one which was small and brown and always friendly.

This is, of course, a gross oversimplification of the work being done, but I think the idea of memory as a collection of attributes rather than detailed sensory records assists in an understanding of how memory can be set adrift. On a purely computational level, changing memory is a matter of switching a few coding units from on to off (and vice versa). Organizational drift can be effected very easily by blanking whole regions of synaptic storage.

How will your mind know? The way it processes data up to the consciousness is purely interpretive; it unpacks a string of binary attributes and assigns sensory details, staples on an emotional reaction, and filters the result through a myopic set of religious and moral definitions. Is there a computational checksum that would reveal tampering? A snapshot of the data set against which the recalled memory is compared? No, I do not think the brain—for all its status as a massive neuro-chemical computer—has much in the way of a useful backup and data integrity verification system. Well, it does: you.

You are the one who finds the deficiencies in memory, the strange details that seem contrary to your world-view, the tiny inconsistencies that are counter to the psychological and physiological rules by which we understand reality. But what recourse do we have to those memories which seem faulty? How can we determine what they should be?

I took Nora to Paris once, building a weekend escape from a memory of a visit when she was in high school. We went to the Musee d'Orsay, and she said the couple in Toulouse-Lautrec's In Bed were us, and the angelic hermit of Redon's Vieillard ailé barbu was how she imagined me when I became an old man and outgrew the hat. I showed her Gustave Moreau's Galatée, and taking my hand, she pulled me into the picture. We made wreathes of red and yellow flowers, and had lunch under the spiny trees. It was, she said, the best time she had ever had in Paris.

There. See? None of it was real in the conventional, in the scientifically observable sense, but what did her brain care? It dutifully stored all of it as strings of binary data sets. Whether she set the switch that identified these data collections as real or imaginary is, well, something only she can say.

I've marked them as true. I mark them all as true. As a data collection that defines personality and identity, is my memory more corrupt or more accurate?