[FRIAM] description - explanation - metaphor - model

[Eric Smith]

I am relieved you brought up the Truth/Power bundling, Glen, because I wanted to but was too much of a coward to do it.

There is a style of speech that I hear often, which goes something like “It doesn’t matter what so-and-so says, or thinks he means.  He is just claiming he owns truth, but I know he is just an entitled group-group-group-assignment, motivated only to exploit or oppress [fill in whoever the good people are].”  My inner translator translates that to my language as “The only thing I care about in life is the fight by which I have constructed my identity, and in my world, there are only two kinds of people: those who are in my army and the enemy.  There are no non-combatants.”   I know my cartoon above is excessive and over-simple, but I may as well admit I have become primed to hear it through time and the accumulation of conflicts, and I can think of a few good exemplars (specific exchanges with specific people over the years) where I think it is fair to say that is really what is there to be heard.  

The problem is, that kind of conduct precludes any other conversation about anything, including most conversations aimed at intellectual clarity, distinctions, etc.  Basically, you can talk to that person if you are talking about or some other way engaged in that person’s fight.  

To me it is not hard to understand that there is a difference between what one is trying to think about, and what one may be motivated to care about.  Certainly, there are some who are so totally consumed by compulsions that they can’t do it ever and so can’t see a distinction, but I think most of us in ordinary life are comfortable with the premise that both can exist, and are capable to some extent of knowing when we drift from one to the other.  Not ideal, and not reliable, but enough that we can see a reason to have both categories.  I assume most postmodern philosophers are complex enough to be capable of parsing such distinctions.  Hence if they choose to entirely conflate them, it feels to me like dishonesty, and often the specific dishonesty of a resentment motive (at the core; it accretes lots of other vanities and problems as it grows institutional.)

This is what I find unpleasant about Rorty.  If he had labeled himself a social critic, I would have been happy to support him (and in that role, I _do_ support much of what he says and I find it insightful and important).  But his delight in hoping he is destroying something that somebody once esteemed (here, the concept of truth, though I have watched him dance like a Stephen King monkey in attacking Weinberg’s efforts to describe some things about how science is practiced) is to me just the posture of the person who is mainly motivated by resentment of whatever he construes as power.  


My comments above are oblique to your main point below about Truth and Power, and the postmoderns being pragmatist, but I think it connects back eventually.  

I have been thinking a bit about pragmatism in the context of a different conversation, which (for reasons not relevant to the thread here) have me thinking there should be a formal version of the pragmatist position that has technical questions in common with ideas we pursue in statistical mechanics, error correction, and things of that kind.  Where I want to get to is that we can all admit to the probable error of all positions on the short term, without concluding thereby that they must reflect claims to power and therefore we can be power-monists, without needing to have both truth and power as primitives.  (I am not branding you as endorsing such a position, but I read you as saying that is where the postmoderns want to be, which is also how I read them).  What I want to claim is that that postmodern position is very far from what I would think of the main conceptual center of pragmatism.

The idea being very lowbrow.  Suppose we are willing to work within the space of concepts and models that physicists have been using for a century, and not worry about deconstructing every word in every sentence in case they might all be hallucinating.  I want to make claims about structure _within_ that space of models and concepts.

We routinely talk about a generating process for some stochastic dynamic, and the process has values for some parameters.  (Rates for a chemical reaction, biases for flipped coins, whatever.).   We then talk of samples from the process, of estimators computed for the samples, and of how the estimators are distributed.  In this lowbrow world, it is unproblematic for a problem with a continuous state space, that a finite sample estimator has measure-zero probability to coincide with the exact value of the parameter in the generating process, but that the generating parameter can still give the value of a stable central tendency for samples.  We care, then, about which estimators are unbiased, which estimation protocols converge with large sample sizes, etc.  All stuff that everybody on this list knows backward and forward.

Things become interesting when there starts to be considerable mechanistic complexity and hierarchy, control relations, feedbacks, etc., so that it becomes _very_ hard to chase through the convergence properties of finite samples.  Hence we see that the biosphere appears to have certain properties stable on geological timescales even though many other things change, but can we justify that impression, or derive from some kind of “first principles” whether a sensible model for the biosphere would be stable in that way?  So far, not.  

The problem of making pragmatism a well-formed position feels like it should have much of that character.  Scientific inference (also everyday inference) is very much “theory-full” in Leslie Valiant’s sense in Probably Approximately Correct.  The theories are controlling systems over how we get rich interpretations from poor observations.  Sometimes the weight of observation can nudge a theory Bayes-wise in a better direction.  Sometimes a bad theory leads to systematic misinterpretation of facts for a very long time (Alchemy, trickle-down, one could go on seemingly forever with examples).  The components have only each other and their couplings with whatever we posit is a “real world” to stabilize them, and whereas we tautologically consider the “real world” to be whatever is consistent by virtue of being what it is, we should take as assumptions that all the components of the interpretive system can be subject to errors in a monstrously more difficult version of the way sample estimators can be wrong.

Biases from unfortunate motives can be one source of sample skew, but that is just one mechanism.  Identifying it, or any other mechanism, seems like a different conceptual problem from trying to figure out what convergence-to-truth can mean in an interpretive system, and to then derive what kinds of properties “truths” can have as the fixed points of such convergences.  For instances, even if I tell you that phase transition theory exists, or that asymptotically reliable error correction exists, you still have the whole scientific domain of understanding how sparse or dense or stable phases can be, how they can be related or interconnected, etc., or what is the domain of applicability of Shannon’s reliable-encoding theorem and how its manifestations vary from context to context.  

It would be appealing to me if some of what we have learned in these much simpler fields (physics of matter, reliable communication) could be bootstrapped into a technical analysis of what pragmatism can be or is.  It also seems to me that there is a kinship between the explanation for the stability (or apparent stability) of very complex things like the biosphere, and the problem of formulating a notion of truth with the right kinds of stability.

To circle back, then, with the complaint that opened this post, when the postmoderns just declare that there isn’t really anything else to think about regarding truth, than their resentments of somebody or some system that they regard as holding power, they make themselves uninteresting for me to invite into my personal world, which has a hard enough time holding together and making any progress on anything as it is.  

Thanks, 

Eric





> On Dec 26, 2019, at 12:25 PM, uǝlƃ ☣ <gepropella at gmail.com> wrote:
> 
> And this is one of the reasons postmodern rhetoric is more pragmatic than modern rhetoric, because it shifts the concern away from Truth and toward Power. It's nothing more nor less than the standard gumshoe technique of following the money. If you want to know why some yahoo said what he said, Truth is irrelevant. What matters is how he might benefit from such expression.
> 
> But many people seem to think postmodern implies a form of pure relativism. I disagree. A postmodernist can still believe in some stably structured reality "out there". But she is willing to employ *both* power-based *and* stability-based analytic tactics.
> 
> A friend recently claimed I wasn't a Platonist because I challenged the idea of a unitary, constant entailment operator (⊢), as well as me claiming that the whole algebra can be arbitrarily changed, at will. So, the question for the Platonist becomes "which parts do we hold constant and which parts vary". I'm still a Platonist ... simply one that's skeptical of anyone's assertion that some part should be held constant/universal.
> 
> As you point out later in your post, of course, we have to doubt our own rhetoric just as much as we doubt others' rhetoric. And that's (obviously) difficult. Personally, posts like this one (https://www.scottaaronson.com/blog/?p=4476) teeter me on a kind of knife edge. It's a great sensation to teeter one way, then another, on some value-based judgement. Did Pinker's tweet provide cover for systemic sexism? It's a kinda Zen Koan ... one of those unanswerable questions whose only proper answer is Mu. But if we look at it through a postmodern lens, Pinker is *clearly* part of the good old boys club ... as crisply a member of that set as Jordan Peterson. He's objectively smart enough to know better than to tweet such nonsense.
> 
> Seth Meyers handles this well with his "Jokes Seth Can't Tell" segments. And the recent Jost/Che bit where they give each other jokes to tell blind handles it well, too: https://youtu.be/Ys786ZsA5tI. In the end, the bane of the rationalists (including Aaronson, Pinker, et al) is their tendency to *avoid* power analytics and focus on truth analytics.
> 
> On 12/24/19 10:08 AM, Steven A Smith wrote:
>> But (BUT) what I think I find disturbing about the truism (oupsie!) that
>> "everything is interpretation" is so often used as the sophists entree
>> into a manipulation, into a switcharoo where the "everything is
>> interpretation" suddenly becomes "let me give you my interpretation in a
>> compelling way that has you acting as if it is somehow 'more true' than
>> the one you started with". 
> -- 
> ☣ uǝlƃ
> 
> ============================================================
> FRIAM Applied Complexity Group listserv
> Meets Fridays 9a-11:30 at cafe at St. John's College
> to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com
> archives back to 2003: http://friam.471366.n2.nabble.com/
> FRIAM-COMIC http://friam-comic.blogspot.com/ by Dr. Strangelove