[FRIAM] Wolpert - discussion thread placeholder

[glen]

There's a nice thread in your responses to 3 of the questions:

5: "alternative SAM"
8: limitation of "mind" but not brain
12: *all* of physical reality

These (and I prolly missed some) allude, I think, to the distinction between the indexicality of the subject (the actual animal-agent) and the linguistic subject (the token "I" when spoken by the animal-agent). In our arguments about an inner life (interior world) versus a behavioristic flattening to monism, we don't articulate well enough that language vs speaker distinction.

Wolpert's trying to discuss built artifacts, whether source code or arts/anthropology *publications* is largely irrelevant, to me. A published paper in an anthropology journal *is* a "finite sequence from a finite alphabet", however distinct it might be from a mathematical proof. So, it's not that hard to widen SAM to include such things. We don't need to write him off as a Scientismist.

And if we do widen the domain of SAM in that way, then we get to a kind of Wittgensteinian [in]effability, even if it's not formalizable like that from Tarski or Gödel. When I hear you poetry fans talk about how poetry gets at some corners/edges of the universe that other forms cannot, what I hear is the idea that poetry brings the linguistic subject *closer* to the ontological subject. It gets into those pathological corners that more explicit artifacts (like prose or math proofs) can't reach. This is especially true with things like cadence, onomatopoeia, etc. *Spoken* poetry is tacitly different from written poetry *because* it is a composition of speaker and language.

I'm not a fan of neuro-linguistic programming. I think it's largely nonsense. But there is *something* that makes it easier to be a guru-in-the-flesh than to be a guru-in-writing. So whatever that thing is, body language, pheromones, whatever, that is also SAM. Prestidigitation is not categorically different than, say, lab chemistry. Our SAM assumes our bodies, in the lab, doing benchwork as much as it assumes computers executing Matlab and hands writing equations.

So, if one buys that rhetoric, Wolpert's mistake is in separating the speaker from the language. There's no problem with the conception of "all of reality". There's only a problem assuming it can be written down (as SteveS mentioned via "not-prestateable"). But to be fair, Wolpert's "Physical Limits of Inference" treats this very issue, which is why I'm pretty sure "What Can We Know" is a bit of a steelman of a position he may not hold himself.

On 10/1/22 14:05, Prof David West wrote:
> continuing in the original thread ...
> 
> Wolpert question 5: my previous arguing that knowledge and information—but of a different order/kind—and "TRUTH" can be found on an LSD trip seems like a negative answer to Wolpert's Fifth. Yes, we do have access to and can learn to use 'alternate states of consciousness' and create/discover alternative SAM.
> 
> skipping six because I am the dumbest computer person in the group.
> 
> Wolpert 7: I am not sure how you would derive a conclusion that human cognitive abilities are constrained by our SAM. First, why the assumption that SAM is the sole apex of human cognitive product? Arts, Anthropology?  I have found a parallel with Wolpert's assumption‚in the work of Ian McGilchrist. The latter argues that our minds and our cognitive abilities "suffer" from the "left brain's limited perceptual and processing mode." The SAM created during a period of left-brain dominance would be constrained accordingly and there would seem to be a correlation: constrained SAM—constrained cognition.
> 
> Wolpert 8:if there is a restriction to finite sequences, then yes, it is a limitation of our "mind" but not our brain. Our brains are massively parallel / distributed processors of massive amounts of sensory input and aggregate, connect, and correlate that data to present an abstracted, simplified, and, in important aspects, imagined REALITY to our mind. Same idea as the originated and perpetuated Maya.
> 
> Wolpert 9: as the least mathematician among you, I will keep my comments as philosophical/speculative as possible. I wrote a long essay on the futility of Software Engineering. In that essay, I coined the term Turing Space,the binary realm of executing programs—the mental model of the state changes of the computer at one step of a program to the next; the mental model the Brooks (No Silver Bullet) stated was beyond human capability to generate/maintain/utilize. My metaphor for Turing Space was the infinite tape in the Turing Machine model. Infinite IS, after all, infinite. There are an infinite number of binary strings that will cause the Turing Machine to start and stop in the exact same state, There are an infinite number of such strings that will do otherwise. There are an infinite number of 'efficient' strings in the infinite set of strings that produce the 'correct' result. There are an infinite number that are 'inefficient'. Software Engineering tells us how to build 
> massive fantastical architectures (systems) in Turing Space. Concern as to the relevance of those structures to anything human is deemed impertinent. But the infinity of infinities of Turing Space, I think, parallels Wolpert's questions and conjectures in #9.
> 
> Wolpert 10: This seems to be just another fromulation of the Anthropocentric Principle: the universe is what it is because that is what humans are capable of understanding. As a firm believer in the possibility of perceiving and utilizing the ineffable, I would have to say no, we can conceive, and even experience, "mathematical" and "physical reality" that are not expressible in terms of what we assume to be our cognitive abilities—primarily language.
> 
> Wolpert 11: anyone who believes in a god / God and in miracles kind of needs to believe in SAMvX.0 are 'beyond' the limitations of our brain. One of the extremely few cultural universals is a belief in the 'supernatural'. So we seem to be able to imagine such constructs even if articulating them is impossible.
> 
> Wolpert 12: Doesn't this require making an assumption that there is some kind of "all of physical reality?" Would this only be possible if physical reality was finite and non-dynamic, i.e. where the All was not constantly variable? if we make such an assumption then we very well might expect our SAM to, eventually, be congruent with Reality.
> 
> davew
> 
> 
> 
> On Mon, Sep 19, 2022, at 2:20 PM, Jon Zingale wrote:
>> What follows are mostly speculations:
>>
>> It is possible that we do not get to have closed cartesianess (with all
>> of its currying and the rest) and so we do not really get to have *all*
>> possible worlds, perhaps only those that are symmetric monoidal. Still,
>> what then does this mean for us, since we can clearly posit cartesian
>> closed categories (like Set) and reason about them. That is, they are
>> somehow afforded to us like any other fiction, and like other fiction,
>> they play a role in our understanding of ourselves (Tennesse Williams)
>> and our understanding of our worlds (Noether[∫]).
>>
>> Glen has me right when he suggests that I am not particularly wed to the
>> idea of a monism; whether it be monotheism, experience, category theory,
>> GUTs, etc... But I do find studying the available monoids to be as
>> fruitful as studying the available groupoids, etc...
>>
>> In Lee Smolin's "Three Roads to Quantum Gravity", he conveys (as Hywel
>> often did) a skepticism toward universal acceptance of the law of
>> conservation, suggesting that a world with clean opposites would be a
>> trivial one. This has me thinking about the role duality plays in
>> modern mathematics (Galois theory, say) where we are not interested in
>> invertible maps between categories with different internal structures
>> (fields versus groups, say), rather we look for best approximations to
>> invertible maps (the adjoint functor perspective). It wouldn't surprise
>> me that that despite the successes of Maxwell to pin down E&M as two
>> faces of the same coin, that our quest for magnetic monopoles will
>> continue to be stymied because the duality isn't exact. That where we
>> attempt to reconcile two "kinds" of things, we will find subtly different,
>> yet corresponding algebras.
>>
>> I mention some of this because duality (and symmetry more generally)
>> may simply be "afforded" to us and not "reality" for us. Still, the world
>> (and I use the term loosely) may reward those that believe (and act on)
>> such a fiction[Ax]. So then, many programs (it seems to me) rely on
>> being able to "dualize" into a larger space of possibilities/fictions,
>> in order to make sense out of what may be much more constrained. It may
>> very well be the case that the world, for instance, *must* be logically
>> consistent and complete and so can only support first order logics, but
>> assuming not, I would feel compelled to ask whether this world was one
>> that has the axiom of choice or not. My intuition (and preference) is to
>> imagine (as Glen suggests) that the in-principle ends of our questionings
>> do not culminate in a single monastic theory ;)
>>
>> At present, I am entertaining Everett's monism, and wondering if all we
>> physically perceive are the moments of decoherence, and that what we
>> experience as particles are little more than the aliasing effects of
>> a wave function shedding its skin.
>>
>> [∫] I am reminded of the Maria La Palme Reyes' (et al) observation in
>> their paper "Reference, Kinds and Predicates":
>>
>> "The role of counterfactual situations in determining the actual is
>> further exemplified in classical Mechanics. To determine the real
>> trajectory of a body, we use the calculus of variations and compute the
>> Lagrangian of all its possible trajectories, most of which are only
>> logically, not physically, possible. We choose as the real trajectory
>> the one for which the Lagrangian has a minimum (or stationary) value.
>> The possible is essential to describe the real."
>>
>> [Ax]. For instance, when chatting with EricS I get the impression that
>> linear classifiers can be unreasonably effective at sorting the bio-
>> chemical world. Despite the improbability of linearly evolving genes,
>> there is clearly a huge benefit in approximating linearity.


-- 
ꙮ Mɥǝu ǝlǝdɥɐuʇs ɟᴉƃɥʇ' ʇɥǝ ƃɹɐss snɟɟǝɹs˙ ꙮ