Slashdot Mirror
New Study Finds People Remember More Than They Think
An anonymous reader writes "A new study has shown that people subconsciously retain information about things they've seen even if they can't consciously remember. From the article: 'Luis Martinez of CSIC- Miguel Hernandez University in Spain and his team "read minds" with the Princess Card Trick, an act invented by magician Henry Hardin in 1905. Participants in the study mentally picked out a playing card from a group of six cards, which then disappeared. When a second group of cards appeared, the researchers had amazingly figured out which card a person had in mind and removed it. Very few people caught the trick: All of the cards in the second set were different, not just the card that people had chosen. This trick is well-known to confuse the masses, even via the Internet a magician's sleight of hand can make it seem as though he/she legitimately "read your mind" A few moments after viewing the two panels of cards, volunteers were asked which of two new cards was present in the first set of cards. None of the volunteers could actually recall which card was present. Despite claiming that they had no idea, when they were forced to choose, people got the right answer around 80 percent of the time. “People say they don’t know, but they do,” Martinez said. “The information is still there, and we can use it unconsciously if we are forced to.”'"
Ok, found it. Neurons operate at 200 Hz, not 10. That gives a brain speed of 24 THz.
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
Can someone please explain the trick to me? Is he picking the right card, or a card that looks like the right card? I mean, if you showed me six cards and I pick one and then you show me a different six cards, I'm going to remember what my card looked like, unless all twelve cards are very similar.
The trick is that the magician, without ever knowing which card you picked, seems to have "magically" taken it out and replaced it with a different card. It relies on the fact that you won't remember the 5 cards you didn't pick, or else you'd notice that all of them were replaced.
However, the point of this study was determining whether you unconsciously did remember which cards were in the first set, even though you could only consciously remember the one you had chosen.
The brain is a machine, so reductionism works just fine. What I did not say, and needs to be taken into account, is that you cannot parallelize a process further than it can be reduced into wholly independent steps. (Interdependent steps should be split into the dependent and independent components, with suitable barrier operations to synchronize them.) Further, any parallel architecture, brain included, is subject to Amdahl's Law.
Computer hardware is capable of matching the human brain today, at least at the level of computation power. You can build a cluster of the required number of nodes, linked together via a hypercube network topology. You'd be bankrupt if you did, but you can do it. Nobody would have the faintest idea of how to program a supercomputer on that scale - you might not have noticed, but parallel programming is a highly arcane art. SIMD is about the only design anyone knows how to program on these proto-Deep Thoughts, but the brain isn't SIMD. It's MIMD. The total number of MIMD engineers out there is less than the total number of Perl 6 gurus. Put them in front of a machine with a few billion nodes and their brains will explode. It'd make a great Halloween video, but it's useless for Strong AI.
Lets say you could find a MIMD guru with the wizardry and dark arts expertise to program where angels fear to tread. Would that match the human brain? Well, still no. We don't have a specification for intelligence and you can't program Strong AI by guesswork alone. Strong AI proponents have tried and it doesn't work.
Ok, let's conjure up a specification. NOW can we match the human brain? Alan Turing proved the answer to that is yes. The brain is a Turing Complete machine, the computer is a Turing Complete machine, either can do the work of the other. You have to allow for the fact that brain cell DNA is self-modifying and that brain wiring is also self-modifying, producing an amazingly powerful and flexible system. You also have to allow for the fact that inter-neuron communication uses analogue or discrete signals, whereas computers are limited to binary, and the brain is incredibly small (reduced distances for signals). A computer with this many nodes would be multiple football stadia in size.
But, yeah, if we could solve the problem of not knowing what the hell intelligence even was, we could build an artificial brain equal to (but slower than) the human brain.
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
The opposite, in fact, is true. Unconscious is actually the correct term, and would be used by educated (at least in psychology) people. Subconscious is imprecise and academically useless, and generally only used in casual conversation, or by pop-psychologists and New Agers.
Excuse me, wtf r u doin?
Guys, I think he's being serious.
I am a scientist (i.e., experimental psychologist) who studies human memory. What is described here is simply the difference between a recall task and a recognition task. Roughly: in a recall task, you have to produce information from memory given some cue; in a recognition task, you are given the information and you have to judge whether it was previously encountered. It is extremely well-know and well-documented in the scientific literature that recognition performance is almost always better than recall performance. In everyday terms, you may not be able to recall the name of a childhood friend, but you may be able to recognize that name among a list of alternatives. The difference between recall and recognition performance is just one kind of demonstration that the entirety of information stored in human memory is indeed much greater than what can be accessed at any given time.