[FRIAM] Doxastic logic - Wikipedia
gⅼеɳ ☣
gepropella at gmail.com
Fri Sep 22 12:00:10 EDT 2017
My answer to Roger's question is "both", FWIW. But my concern seems slightly different from both Marcus' and Nick's answers. I'm more concerned with the granularity of the updates/iota. Nick's 70/30-clean/scramble is pretty fscking coarse. As I said early on, my beliefs/skepticism is *never* that coarse. Marcus' set of equivalent solutions gets closer to what I care about... a kind of measure of how many options one has at any given *instant* in the action process. And I also care about the boundary of that set. Which course corrections can I make that still lead to a satisficing objective (like crashing my bike without brain damage), which lead to failure (brain damage), which lead to optimal outcome (dodging the left-turning old lady completely), etc.
I maintain that some of this complicated problem solving is conscious and some is subconscious (muscle memory as well as the lizard brain). And I tend to believe that the spectrum between the two is fine-grained. I.e. there is no disjoint, binary, distinction between "things I do with full belief" versus "things I (don't) do with full skepticism."
On 09/22/2017 08:26 AM, Nick Thompson wrote:
> So I leap across the chasm believing that I have a 70 percent chance of making the jump but knowing that I have a 30 percent chance of not making it. I think James would argue that to the extent that one paid attention to the 30 percent, it is actually increased. I.E., if you jump ambivalently, you are less likely to make the jump. And that would be because an ambivalent jump is functionally different from a confident one. For instance, to the extent that you prepare yourself to grab at the cliff as you miss, you ill-prepare yourself to make the jump cleanly.
On 09/22/2017 08:31 AM, Marcus Daniels wrote:
> Underlying such a network is some generating process, and the belief is about that ongoing process, as tabulated by joint and conditional probabilities. Some of the imagined degrees of freedom may not be relevant in an applied setting (e.g. pilot waves or a multiverse) and are acceptable reasons for having probabilities, but others can and should be explained by hidden or external variables. The more these variables are made explicit, the more precise and falsifiable the predictions can be. Ideally, one would have a network of logical predicates that deterministically lead to one or a degenerate set of equivalent solutions.
--
☣ gⅼеɳ
More information about the Friam
mailing list