Math4Wisdom

M4W spaces

Investigations

Tools

Topics

Questions

Other Notes

Explore

Topics

Surprise

⁠

A taxonomy of surprise definitions⁠

—

https://youtu.be/zICcGukKbAA⁠

⁠

https://www.sciencedirect.com/science/article/pii/S0022249622000517⁠

⁠

Surprising!

⁠

KU: having immersed myself in the language of

Active Inference⁠

(an ongoing process) I’m into generating

text blobs⁠

(such as what follows) wherein I work to achieve a “Vulcan mind meld” twixt the Active Inference and Synergetics namespaces. [

blogged version⁠

]

⁠

Synergetics⁠

is a namespace used by “anticipatory design scientists” (cliche phrase) to explore a geometry of thinking.

Synergy (the concept) suggests unanticipated (surprising) developments when the (perhaps already analyzed) parts cohere and inhere in the form of newly emergent wholes, meaning the unexpected is to be expected, if not precisely.

Chance and randomness play a role, sometimes for the better (luck, windfall, boost) leading to new scientific discoveries (e.g. radio, penicillin) that transform our continuation strategies. A good science lab is a studio within which to produce and confirm surprising (novel) findings.

Between an inside (concave) and outside (convex) comes our Markov blanket membrane, our

omnidirectional halo⁠

model, our system of relevance between twilight zones, that takes its own manifestation, its own existence, as further evidence if its suitability as a model (life form).

⁠

from RBF’s No More Secondhand God (1963)

from RBF’s No More Secondhand God (1963)

⁠

We may always recurse to say a model itself relies on models. A model is a set of biases, preferences, simulations, generalizations, generative vectors, forming a semantic alchemical space, internally to a containing model. Models have a natural half-life, a decay curve, unless updated, presuming a continually changing environment.

Sense decays towards nonsense and obsolescence unless actively tuned and maintained.

Our system’s context might be “the village” (township, base, outpost) organized around various types of seer, chief or shaman, each of whom internalizes (models) the village, its procedures and ceremonies, rules of government, and its context (the rest of the world). Their collective steering produces

synergetic results⁠

In general systems theory (

GST⁠

), the personal workspace (PWS) may be variously interpreted, as a lab or studio, game pod or cubicle, an individual on the beach (with a dog maybe).

Synergy comes from the “value added” i.e. between the energy in (including sensory input, food calories, art supplies) and energy out (including action, making, crafting, cooking), i.e. from the gap between energy in and energy out filled by work (personal activity) of whatever nature.

Work is potentially value adding (varying criteria apply). Energy throughput may also be entropic, as when a dam or bridge gives way, and no longer serves a purpose.

What would be the consequences of sometimes swapping out “surprise” for “dissatisfaction” i.e. what drives a model is to minimize dissatisfaction and maximize satisfaction with itself, as a model, based on confirming evidence.

“The model is satisfactory, captain”. If the model fails to account for too much, there’s this audible buzz of uncertainty (a sign of free energy) and a sense of incompleteness or falling short, and perhaps a consequent need to wake up (increase awareness and alertness).

Couldn’t this be accomplished with no change at all to the math? It’s only the nomenclature that’s shifting here.

Perhaps the goal of a model is to become tautologous meaning it merges with its own implementing grammar, thereby becoming self evidently true insofar as it supplies its own context. Easier said than done. Not every theory is able to become practically unfalsifiable and yet non-trivial as a source of new cognitive systems, heuristics and gestalts, much less remain in fashion in a competitive landscape.

We allow that life forms are “true” in the sense of “ongoingly operational from day to day” i.e. functional. “The proof is in the pudding”. Crabs and penguins are model creatures (agents) with positive prospects (such as the ability to reproduce) in a given environment. We need not rank them in terms of which is “better” in the sense of “more true” on a planet where both are thriving. Ditto thriving cultures.

One might think crabs are less beautiful or cute than penguins, but then why clutter one’s thoughts with such spurious judgments and preferences? When does one’s own train of thought turn to “mere noise” in one’s own estimation i.e. how tightly should one hold to random expressions opinion, just because they flicker by? “Who made you the judge?”

How closely does one identify with a model? Another parameter. A trusted friend gives me their best interpretation of what’s going on in a bar scene I’m about to enter, leading to me trying to test the model knowing it’s incomplete and that I should be ready to drop it for a better one. I treat this friend’s model as highly tentative, until I get a lot more evidence one way or another. Models come and go.

A key question is how to model and/or talk about switching models and/or constructing better ones even while getting value from the current one. Sometimes a model is recognized from the get go as bad in the sense of insufficient, and the point of the exercises which follow may involve abandoning it for another one, as when we speak in terms of “jumping ship” and/or “leaving a sinking ship” (e.g. saying goodbye to a childhood belief system centered around Santa Claus).

Other times a model is recognized as “serviceable” but in need of various forms of maintenance. When you actually visit a place you’ve only heard about and imagined, that may involve valuable updates to the mental models in question. Learning about the world from one’s armchair, versus touring it in person, results in models of different character. Both may be satisfying in their own way.

Kierkegaard: to will the good is to will one thing. Models are about simulating possibilities and choosing those scenarios that seem more likely to work out in terms of achieving a model’s goals. Where do these goals come from?

The value of “personal freedom” is about the freedom to set one’s own goals, sometimes in perceived conflict with the goals of others, but also with the potential for collaboration as a strategy.

The first step of an agent contemplating next moves may involve considering peer agents on the team and their specific capabilities. Or does one’s model show one is peerless? When classifying models, one might ask: “how me-against-the-world is it, on a scale of 1 to 10?”. Complementary question: how hostile to the self is one’s environment? Models might be modeled as outfits (blankets, outfitted avatars).

It often takes a 3rd party to espy key divergences between a model and its territory. Error correction depends on others’ perspectives, more often than not. We factor in the judgements and beliefs of “those in a better position to know” which in itself is a changing picture (parents, grownups, adult voices on the radio, preacher...). The quest for “trusted sources” is properly endless provided one’s sense of curiosity is likewise boundless.

In reality, changing one’s model may be no easy task because with models come habits (including habits of thought) such as always and only saying “10 squared” when reading 10 followed by a 2 superscript. Some people might even picture a 10-by-10 square (or picture picturing it), not that the presence of such qualia is necessary to the communication. So why not “10 triangled”? Few would ever think to pause at such a question. Countering habits involves inserting pauses.

People go to camps, farms, schools, conferences, special events, to develop new physical and mental habits while extinguishing unwanted ones. Actors grow into new characters, based on a script, storyboard, reading rehearsals. One has to imagine oneself into a role, and that takes time, especially when the role itself may not be clearly articulable pending more information.

How does one prepare for one’s next role on the world stage when the action seems so chaotic? That’s where your model comes in, perhaps with refreshing counter questions and advice such as: who else do I know looking for an active part in World Game? Start with what you’ve got. Another question: who says you need a new role? Just asking.

Now it sounds like my chatterbox has been training in the self help section of the bookstore. True enough: it has read pretty widely, and lots of modeling has gone on. The large language model brain is regeneratively thought provoking.

The quadpod is an image of a system maintaining an equilibrium position in an overall phase space, meaning a net zero change in the model as long as it proves satisfying. Fine tuning is always needed and the energy budget that derives from being off or wrong gets used to then close the gap, and so on.

Modelers might ascribe properties to the four canonical directions, the tradeoffs. Four superpowers. Each one used alone serves to counter the other three used together.

⁠

Load content from www.math4wisdom.com?

Loading external content may reveal information to 3rd parties. Learn more

Allow

Want to print your doc?
This is not the way.

Try clicking the ⋯ next to your doc name or using a keyboard shortcut (

CtrlP

) instead.