Thinking about thinking
"Consciousness - our brain contains a virtual serial machine built on top of the hard wired evolved natural predominant parallel processing."
I like this quote from learningevolves, it encapsulates the higher order thinking that is missing from Connectivism. The virtual machine uses symbol systems, often but not always language. The symbols can be visual images too but they do run serially.
Visual thinking can consist of constructing a mental model or simulation and running that simulation in our head. The simulation runs serial mode. We observe this thought experiment and then use language to extract conclusions.
By comparison, making connections and data retrieval run parallel.
External storage, be it pencil and paper or delicious bookmarks or the scrabble letter rack expand our limited working memory and assist both our parallel and serial processing.
I once had a dumb cow and a smart cow. If I stood on the other side of the fence with a bale of hay, the dumb (Hereford) one would move as close as possible and then stand and bellow.
The smart one (Jersey) would process for a few seconds then run away from the hay, out through the gate and back. Was it making connections, running back through all the stored chess moves triggered by fence and hay till it retrieved a success scenario? Was it running an A* path finder algorithm on a visual map model in its brain?
Bad navigators use a 1D map representation in their head, probably stored as linguistic symbols "first right then turn left at the shops, first right at the railway line" Good navigators store a visual representation of a 2D map. Probably both run in the virtual serial machine.
Then there's autopilot mode. Ever ended at home when you intended to go somewhere else? The good old Connectivist parallel processor recognised landmarks along the way and fired up the drive home connections.
I like this quote from learningevolves, it encapsulates the higher order thinking that is missing from Connectivism. The virtual machine uses symbol systems, often but not always language. The symbols can be visual images too but they do run serially.
Visual thinking can consist of constructing a mental model or simulation and running that simulation in our head. The simulation runs serial mode. We observe this thought experiment and then use language to extract conclusions.
By comparison, making connections and data retrieval run parallel.
External storage, be it pencil and paper or delicious bookmarks or the scrabble letter rack expand our limited working memory and assist both our parallel and serial processing.
I once had a dumb cow and a smart cow. If I stood on the other side of the fence with a bale of hay, the dumb (Hereford) one would move as close as possible and then stand and bellow.
The smart one (Jersey) would process for a few seconds then run away from the hay, out through the gate and back. Was it making connections, running back through all the stored chess moves triggered by fence and hay till it retrieved a success scenario? Was it running an A* path finder algorithm on a visual map model in its brain?
Bad navigators use a 1D map representation in their head, probably stored as linguistic symbols "first right then turn left at the shops, first right at the railway line" Good navigators store a visual representation of a 2D map. Probably both run in the virtual serial machine.
Then there's autopilot mode. Ever ended at home when you intended to go somewhere else? The good old Connectivist parallel processor recognised landmarks along the way and fired up the drive home connections.
Labels: connectivism, constructivism, OCC2007, semiotics
posted by Unknown at 8:11 AM
2 Comments:
try this one on your friends, the parallel processor gets fooled every time, ask these questions out loud of another:
spell hop?
spell shop?
what do you do at a green traffic light?
Tony, consciousness is a tricky topic: the epiphenomena of our serial experience from a demonstrably parallel architecture. My take from a PhD in Cog Sci in the connectionist camp is not that the serial machine sits atop the architecture, but that the linear 'story' is emergent from the architecture. But it's still an ongoing debate. Clearly it's got to be the trace of the associative patterns of activation...but who's looking at it? Homunculi...
Of course, I'm still looking for the real 'take home' message from the neural net stuff! What do we do differently in our learning? I think the social constructivists and cognitive apprenticeship still captures the necessary implications. But a fair go at the connectivists, I reckon.
Post a Comment
<< Home