[FRIAM] falsifying the lost opportunity updating mechanism for free will

[∄ uǝlƃ]

OK. I agree that feedback need not always be higher order than the process being controlled. But that's what I was positing, that a composer and editor *are* higher order than the processes being composed and edited. For some reason I can't articulate, I don't think the *thin* mechanism you're providing will allow for the lost opportunity sense I'm targeting.

Your mechanism seems too thin, to me, because it relies mostly on stigmergy and a kindasorta Markov property. I think the memory and editing, to properly cover the as-yet-vague phenomenon we're targeting, needs a parameterization that allows for playing around with [in|de]creasing the memory and [in|de]creasing the composition scope (both [co]domain). E.g. I'd like to be able to set the memory size to something like 100 years ... so that an old person can consider the lost opportunities of, say, being born a bastard ... which would be something one has little control over. Then shorten the memory to, say, the last time you smoked weed and told your friend that they're an uptight priss and you hate hangin' out with them ... something you might [†] be able to modify in a week or two but is largely unrelated to how well you trade stocks at your day job.


[†] Again, no assumption of [non]determinism is being made, here.

p.s. And none of this need be abstract, either. Long-scope feedback loops can be concrete.

On 6/22/20 9:54 AM, Jon Zingale wrote:
> Glen: "2) dampening edits or negative reinforcement"
> 
> I am positing that negative feedback also exists in such a system and is
> partly responsible for the phenomenon of delta switching. Different flows
> in branches can be construed as probabilities, but with the additional
> action
> of sedimentary deposition, which can supply a dampening mechanism over
> branches. Perhaps the key to completing the picture is to note the potential
> between the source and mouth of the river is effectively constant.
> The sedimentary deposition has a similar effect on the flow that memristors
> have in physical neural nets. I guess my point is that *feedback loops*
> need not be understood in the abstract as explicit loops on a graph.
> The materials and the physics have the property without the river's flow
> being explicitly a loop.


-- 
☣ uǝlƃ