The large rectangle is a sheet of blue cortex containing pink processes, with most of the process structure being expressed two dimensionally, but with the possibility of overlap and overlay.
Some of these processes will only exist in the sheet and make no connections either in or out. But, other things being equal, processes taking input from the senses will have more power than those that do not. Nothing like a bit of real input to get the brain going.
But processes will have exclusive access to neither input nor output. We might have, for example, several processes trying to connect to the motor systems for the left leg and we will need some kind of conflict resolution machinery. Perhaps something as simple as winner take all. And perhaps something else to make sure that the same process gets control of both legs – at least nearly all of the time. There will be at least some of the control machinery deployed in relational databases such as Oracle – but it seems very unlikely that things will be as cut and dried as they are there. Two-phase commit protocol not (see reference 2).
These processes will have a life of their own and they will have a life cycle. They might replicate. They might be spawned from some stored template. They might wax and wane, they might expand and contract. Powerful processes might take space (and resources) from neighbouring but weaker processes. Processes will also be subject to outside influences: the host environment might, for example, be able to simply switch one off, in the way of Task Manager on a Windows PC. Or perhaps not. Perhaps processes in brains, like processes on computers, will be able to run out of control.
Processes will have power, a positive real number which expresses, in some way, how much is going on – although I am not yet clear whether this is an input or an output – or perhaps a bit of both? Perhaps processes will have power at a point, with the power of the process as a whole being the integral over the area of the process. We might say that the area of the process rectangle is 1 and the maximum value of point power is 1, giving a maximum possible process power of 1, only reached by a single process taking all the available resources. We might also measure power in from sensory areas and power out to motor areas.
Some of these processes will be conscious and I need to think a bit more about how the threads, the several processes – maybe one for touch, one for hearing, one for one’s sense of breathing and several for sight – making up any particular conscious experience need to be related, need to be joined together. Does one need to align them next to each other, rather in the way that the long strands of complex proteins get lined up for the purposes of productive interaction? Does one lie one on top of the other? Or do we just need sufficient overlap, for example of processes B and C above? Or is the simple geometry suggested by the diagram not helpful at all?
Conscious processes will have significant power, although the relationship between power and consciousness is more complicated than just saying that the process is conscious if it has enough power, if power is greater than some constant or other. But there might be a necessary but not sufficient condition of that sort.
I also need to think about how the processes talk to the global workspace, which I see as a place where processes can talk to each other, where data can be stored for a while. But my thought is that it is quite a small space, used for sharing data about at the level of the word. So there might be a bit of global workspace saying that a certain horse is the present object of attention. And perhaps the space as a whole can be thought of as a character string, formatted as a bracketed, labelled list, rather in the way of html. For example: ‘workspace(1=horse(brown tail=long) 2=six colour=red tone=bad pain=severe)’.
Reference 1: http://psmv2.blogspot.co.uk/2015/08/its-dogs-life.html.
Reference 2: https://en.wikipedia.org/wiki/Two-phase_commit_protocol.
No comments:
Post a Comment