[FRIAM] The epiphenomenality relation
thompnickson2 at gmail.com
thompnickson2 at gmail.com
Thu Dec 2 12:35:44 EST 2021
Frank,
Attached is the article by Sober on causal forks. Sober is a Good Guy in general, but this text seems weak. I am sending it along again because I still don’t really understand it and I am directing it to you because it probably falls to you to explain it to me. I put it on the list because it just might be relevant to the ongoing discussion.
My problems arise from the concept of “screening off”. I have tried to translate the probability notation into words and I get something strange like,
The probability that I answer the phone given that it rang is the same as the probability that I answer the phone given that it rang and given that somebody dialed it.
As I understand math, this will be true when and only when somebody has dialed the phone. As I understand language, the statement is a truism because the phone will never ring unless it has been dialed
His argument, later in the piece, that the postulation of unknowns that stand between stimuli and multiple responses is heuristic because it might suggest other variables that “collide” with known stimuli to modify responses, is entirely reduceable, as Skinner would assert, to the statement that there might be other observables that effect the relationship between known observables – i.e., no inner postulations necessary. But remember, I stipulate that if your research interests are physiological, then, of course, postulations of physiological mechanisms is useful.
Note that Sober stipulates that his believe that that the “intervening” variables are potent has nothing to with our (ie, yours and mine) old argument about whether we have privileged access to our own inner states.
As to the epiphenomenality thread, I am beginning to think I am not ready to write the paper. This has nothing to do with whether I am an expert in the area, but only to do with an increasing sense that I don’t know what the hell I want to say.
Thanks for taking the time, if indeed you have the time to take.
N
Nick Thompson
<mailto:ThompNickSon2 at gmail.com> ThompNickSon2 at gmail.com
<https://wordpress.clarku.edu/nthompson/> https://wordpress.clarku.edu/nthompson/
From: Friam <friam-bounces at redfish.com> On Behalf Of Steve Smith
Sent: Thursday, December 2, 2021 10:21 AM
To: friam at redfish.com
Subject: Re: [FRIAM] The epiphenomenality relation
EricC/Glen -
I'm glad we agree. I made the same points here:
https://redfish.com/pipermail/friam_redfish.com/2021-November/090981.html
https://redfish.com/pipermail/friam_redfish.com/2021-November/090983.html
To reiterate, we can't reverse engineer a builder's intention from the artifact.
We can't mind read (even our own).
To go even further, we can't even do a *complete* job of characterizing the aspects of a thing, the aspects of environments, or the relations between them.
All models are wrong (though some may be useful).
Parallax is needed across all scales and in both directions. Polyphenism is parallax on the thing. Robustness is parallax on the environment. And counterfactuals are parallax on their coupling.
All systems (existing within the same light-cone) are "nearly decomposable" ?
Herb Simon Sez: https://www.jstor.org/stable/1909285
One of the attractive qualities of modal realism is that it addresses both consistency (through concrete possible worlds) and completeness (through counterpart theory) in positing and testing various models. The problem becomes one of discovering which world you inhabit *from the data*, not from whatever abstracted models you may prefer.
Lewis's Modal Realism <https://en.wikipedia.org/wiki/Modal_realism> is a new one on me, but very interesting framing. Only skimming the Wikipedia Article on the topic leaves me with only enough information to be dangerous... so I am refraining from rattling on about all of my reactions to it's implications (for me) and in particular some of the objections listed there to his theory. From this thin introduction I think I find Yagasawa's extension of possible worlds being distributed on a modal dimension rather than isolated space-time structures (yet) more compelling/useful?
And what would Candide <https://en.wikipedia.org/wiki/Bildungsroman> have to say about this?
On 12/1/21 6:35 PM, Eric Charles wrote:
Me -> We've imputed in all cases. Certainly we can assume artificial systems were designed for a purpose, but we still don't know what that purpose is without imputing a model onto that system. And, in both cases, we could proceed to experiment with the system, in order to test the predictions of the imputed model and increase our confidence that we have imputed correctly. The ability to do these things does not distinguish between the two types of system. There are long and respected scientific traditions using experimental methods to gain confidence in our understanding of why certain systems were favored by natural selection, i.e., to determine the manner in which they help the organism better fit its environment.
Me -> Well.... it might be reification in some sense, but that term usually implies inaccuracy, which we cannot know in this case without experimentation. Even with a system we designed ourselves, where we might have a lot of insight into why we designed the system the way we did, we certainly don't have perfect knowledge. All we have there is a model of our own behavior to impute off of. Once again, this doesn't clearly differentiate the two situations. In all of these situations it is a mistake to uncritically reify our initial intuitions about the system's purpose.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://redfish.com/pipermail/friam_redfish.com/attachments/20211202/10fb8dbb/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Sober Forks no comments.pdf
Type: application/pdf
Size: 635069 bytes
Desc: not available
URL: <http://redfish.com/pipermail/friam_redfish.com/attachments/20211202/10fb8dbb/attachment-0001.pdf>
More information about the Friam
mailing list