The differences between "wrong"; "incorrect"; and "limited"...

[Judith Rosen <***>]

Dan brought up something that, upon further reflection, actually presents an interesting conundrum: How do we define the word "wrong"?

My definition has always been: If some "thing" (where "thing" means some answer, explanation, model, etc.) is fundamentally inapplicable/inaccurate for that which it is supposed to represent/pertain to, then it's wrong, even if it seems applicable on the surface, under certain circumstances (turn it sideways and squint!). Ultimately, this is also the difference between a simulacrum and a true model. (Side note: sometimes the line between models and simulacra can be a bit blurry. I attribute this to the fact that there are always going to be lousy modelers who think like simulacrum-builders.)

I suppose a more user-friendly term would be "incorrect" rather than wrong. "Wrong" tends to invoke moral issues and personal behavior judgments, etc, doesn't it-- as in "knowing the difference between right and wrong"... So I will switch to "correct and incorrect".

Dan asked:
maybe we have to be more
specific about right and wrong, maybe by citing utility
or functional value or cost/benefit, etc.?

I agree, obviously, with the first part or I wouldn't be writing this. (Thanks, Dan.) Perhaps whether something is absolutely correct or only partly correct isn't the only distinction between correct and incorrect... Perhaps it's only a distinction between correct and correct-enough-under-THESE-(specific)-circumstances. However, that's really a different distinction, isn't it? To me, it looks less like an assessment of the model's entailment pattern for accuracy, compared to the system it supposedly represents... and more like a distinction over the correct/incorrect-ness of using that model to make decisions which will affect the actual system. So perhaps these are two separate areas of concern, and it would require some serious relational thought capability and good judgment to discern First: the correctness of the model's applicability and, Second: in the case where the model is only partly "correct" in its entailment relations to the system being modeled, when do those circumstances exist such that it is "correct-enough" to use. (In my view, additionally, it would always be wise to be aware/mindful of the limitations AND the reasons for the limitations-- which is apparently not the case, currently, with state-based models of reality in physics.) Be that as it may, it's clear that in order to make an intelligent assessment about the limitations of the model and about when those circumstances exist that the model is safe to use, one would have to do some serious relational analysis, right? Which brings us, then, to the suggestions Dan offered for basing a decision.

There are three possible criteria listed, for judging the correctness or incorrectness of some scientific model in any given situation: Utility (can it be useful?); Functional Value (how well will it work for us?); and Cost/Benefit (this is a loaded one! It could refer to many different ratios and some of them, I have to admit, are necessary. Even critically important. However, there are some I've seen people use in this category as a specious justification, using a very short timeline to derive their numbers or some other trick of expedience, and THAT I consider to be morally "wrong".) After turning these ideas over in my mind, I have concluded that this list of criteria is dangerously incomplete if it were used the way it is (which is not a criticism of Dan-- he was just tossing ideas out here, off the top of his head, and I'm glad he did). For example, I don't think we could accurately judge or arrive at an answer for any of those three questions unless we do some other assessments, first. So, to this list, right at number one, I would add: How correct have the modelers gotten the entailments of their model? That's a biggie, and it requires periodic reassessment over time, as our technological capabilities progress, giving us access to more of any system's underlying entailment, etc.  Then, at number two, I would suggest: If the model commutes only partially, where are the entailments off? (The answers to this will become very important in any cost/benefit analysis we do later.) Again, we should periodically reassess this aspect with our current capabilities-- not just assume that because it once seemed to commute really well that it still does (It may never have fully commuted at all, only partially, under certain circumstances.....) Similarly, we will need answers to my next suggested addition: 3.a.) Will the incorrect aspects of our model's entailment structure affect our project under consideration and/or the system the model refers to--  and, if so, how? A corollary to this is: 3.b.) What other systems are affected in relation to any affect on this one?  In other words-- number three basically asks: What will be the side effects of using the model in "this" project, with the entailment structure of the model limited in its correctness in "that" way?

My concern over all of the above is that we (humanity) don't even recognize relational causality as a player, in any scientific way. That was my father's whole gig, frankly: that We need to. In order to predict what kind of side effects eventually come from what kind of entailment dis-relation in a model compared with the system it models.... we really need to study the relational aspects of side effects in a systematic way. As far as I know, no branch of science is involved in doing that. Does anyone on the list know of one?

Judith

Web address: http://www.rosen-enterprises.com
BioTheory: An electronic journal of general science based on the Relational (Rosennean) Complexity Paradigm