Determinism Doesn’t Mean a Micromanaging God

“Durdle Dwarves” is a simulation where little “Dwarf” pixels dig-through and build rock-like structures. They do this according to a set of 20 rules. A rule tests a Dwarf’s vicinity and provides a response. (For example, if a Dwarf notices a drop-off to his left, he’ll build a new piece of “bridge”; that’s 1 rule of the 20.)

The simulation is 100% deterministic, and the state of the world at a certain time is a strict function of those initial rules, plus the starting state.

For instance, this starting state:


Yields this world state after 13,500 ticks:


Let’s pretend we don’t like something about this result. Perhaps that structure just below the middle — the one that looks a bit like a road with a lane divider — displeases us.

We could just pause the simulation and erase it, of course.


Because we have the arbitrary power to change things at any time, you and I are sovereign over this world.

But each interruption like this comes at a cost. Even though we were displeased by the road-like shape, we’re also displeased to intrude upon the natural flow of events. In our set of innate interests, one interest is that the world have as few of these intrusions as feasible.

Another option is to change the starting state of the world. Let’s make our initial two “Adam & Eve Dwarves” one block farther apart.


And here’s what things look like after 13,500 ticks:


Hooray! We got rid of that nasty road-like pattern.

But there was a cost here, too. The world is dramatically different now. Here’s the “Before & After”:

This kind of thing is nicknamed the “Butterfly Effect” in chaos theory. The tiniest of changes, over time, produces explosive world differences in nonlinear and/or interactive systems.

This is irritating, too, because we liked the world basically how it was, we just didn’t like the road pattern. Now everything is way, way different.

How annoying!

Let’s define micromanagement as, “Controlling things with surgical precision.” Is there a way to micromanage-away that road, so that we keep the rest (through our precise surgery), but also without using the blunt force of the eraser?

You might be thinking, “Why don’t we alter the starting rules, so that the road never appears?”

Unfortunately, merely altering one of the 20 core rules, or adding just a few rules, has the same explosive “Butterfly Effect” that dramatically changes the world state at frame 13,500. So, this would fail to micromanage.

As it turns out, the only way to pull this off is to have tons and tons and tons of core rules.

And this, of course, has the cost of being horribly ugly and inelegant.

Some folks would tolerate that inelegance. But you and I agree that this basically defeats the point of creating these guys at all. We take pleasure — let’s say — in the emergence of these forms out of an elegant, simple foundation.

Where This Leaves Us

“Durdle Dwarves” is just a computer program.

However, it’s such an obvious deterministic cloister that it serves as the “worst case” for verifying whether determinism always means full micromanagement.

The answer? It doesn’t.

If we have an innate interest in maintaining systemic elegance, then the degree to which the world is micromanaged is equal to the degree to which we have found it warranted, in certain rare circumstances, to intrude, even as we’d rather not intrude.

Notice that our innate interests have what’s called “circumstantial incommensurability” at the “13,500 circumstance.”

It’s not power-weakness that makes a wholly-satisfying solution impossible. Remember, we’re 100% powerful over that world.

Rather, a wholly-satisfying solution is arithmetic nonsense.

What This Shows Us

An entity can be 100% sovereign and powerful over an interactive deterministic world, including that world’s core rules, starting state, and modifications to any subsequent state.

But if that entity has innate interests in an elegant ruleset, then the world is not necessarily micromanaged. Then, if the entity never intrudes, then the world is not micromanaged at all. And if the entity intrudes to some limited degree, then the world is micromanaged only to that degree.

Primary Causation & Secondary Causation

If we start with an elegant core ruleset in a sizeable interactive system, then as time goes on, I will lose surgical control over that system, even if I’m 100% sovereign over it.

True, but pretty dang counterintuitive.

It is only when I find it justifiable to intrude that I can wrest-back some amount of surgical control, temporarily.

  • The reductive perspective is that since these decisions are all “up to me,” everything in the system reduces to “my causation.”

    This is the perspective from which St. Irenaeus wrote in the 2nd century, “The will and the energy of God is the effective and foreseeing cause of every time and place and age, and of every nature.”

But this fails to capture something important to me, doesn’t it? Indeed, a formative perspective is needed to capture my interests.

  • The formative perspective is that there is a meaningful — according to my interests — difference between primary causation, the stuff that glows brightly with my exceptional exertions of power, and secondary causation, the “teleologically-dimmer” stuff that starts to act weirder and weirder as time goes on.

    This is the perspective whereby we rightly deny that God “authors” sin under chaotic determinism; rather, he broadly suffers it.

These two perspectives are simultaneously true in heterophroneo (a concept we’ve been exploring on this site that’s similar to Aristotle’s hylomorphism, but not exactly).


  • Those who say “under determinism, primary and secondary causation have no meaningful distinction” are mistaken.
  • Those who say “under determinism, a sovereign entity has 100% micromanagement, irrespective of that entity’s interests” are mistaken.

Please see last year’s article, “The Sun Also Rises,” to explore other philosophical and theological puzzles that “form heterophroneo” finally puts to bed.

Explore “Durdle Dwarves” for yourself here. Press “Start” to see the world erupt chaotically but deterministically, according to a small set of rules.

Short and sharable:chaos_cheat_sheet


About stanrock

Husband, father. Professional game developer, software engineer, & social product analyst. Theology debugger. Fun theology experiments at

11 responses to “Determinism Doesn’t Mean a Micromanaging God”

  1. Josiah Henderson says :

    > It’s not power-weakness that makes a wholly-satisfying solution impossible.

    But isn’t it “knowledge-weakness” then? I don’t think it’s arithmetically impossible to produce an “elegant” ruleset and starting state that will produce fig. 4 (the world without the road we don’t like).

    • stanrock says :

      A ruleset tailored to make fig. 4 would need to be *radically* tailored for a *single* timeslice. I want to produce fig. 4 while not drastically changing the rules that pleased me otherwise at innumerable other timeslices.

      • Ryan says :

        Four years late to the party, but I just came across this and had a similar confusion to Josiah’s. I’ll take a shot at it anyway 🙂

        For a person (let’s call him Hum) to tweak this ruleset to prevent roads entirely, it would take a stroke of mathematical genius to design a single, elegant “rule 21” that would prevent this universally without changing any features around it.

        There is a trivial solution: “If road structure would exist as the result of my action, do something else”. This solution requires the following: a definition of what constitutes a “road” (which I’ll assume exists, at least in Hum’s head, since Hum must know the construct of a road which he is trying to avoid), a method for reading much more data than simply the local space around him (since whether an action “creates a road” or not could depend on the data in spaces arbitrarily far away, depending on that definition), and an appropriate action to take (or not take) to prevent the construction of the road. As far as I can tell, these rules would make it impossible for a dwarf to construct a road.

        However, this trivial solution (like most) comes at a cost: it requires a possibly infinitely complex definition of “road” (although, this seems to be given from the premise), it requires non-local information (in fact, depending on the definition, it could require ALL the information in the universe/game), and would require an extremely complex algorithm to re-route the dwarf’s behavior if it’s determined that the intended action would create a road, such that it preserves whatever properties Hum was pleased by in the initial ruleset.

        Hum, reasonably, determines that this would be impossible. He cannot formulate his concept of a “road” into precise enough terms within the language of the ruleset, the computational complexity of gathering data at each time-slice and then determining if it is a “road” is too great, and even if those issues were solved, the mere butterfly effect at play in this chaotic system would ruin his previously perfect (aside from the roads) simulation. He gives up.

        However, his friend (named Ged), overhears him complaining about the infeasibility of the project and sets out to prove him wrong. Ged is a programmer like Hum, but far surpasses him (infinitely so, even!) in every way. Ged knows exactly what a “road” is, and has developed a ruleset language of his own that allows him to express this with infinite precision. Additionally, Ged has built his own computer, one which has the seemingly magicalability to perform an infinite amount of operations in any finite amount of time. Where Hum had an issue with finding an “elegant” way to prevent roads, Ged can ignore this criterion completely: with his infinitely fast typing speed, unlimited storage space, and infinite intelligence, he has no problem specifying an infinitely complex “rule 21” to prevent roads while keeping everything else the same. In fact, he may choose to split this up into n rules 21 through 21+n, since he is capable of modeling the interplay of multiple rules in his head with infinite precision and infinite speed.

        Hum, at their next get-together, complains once again about how the only way to remove these infernal “roads” from his beautiful dwarf simulation is to manually remove the road, ruining the fun of “writing the rules once and watching the beauty unfold”. Ged, with more confidence now, asserts that it IS possible to prevent roads using only starting conditions, and he CAN maintain everything else Hum liked about the original.

        Ged shows Hum the solution, but Hum isn’t satisfied. He says that while it fulfills everything he wanted, it isn’t “elegant” enough. Ged either added too many simple, interacting rules, or he added a single rule that was far too complex for Hum to consider it “elegant”. To Ged, with his infinite mind, this makes no sense: he doesn’t see things in terms of complexity and simplicity (constructs which are helpful for finite, human-like minds), and thus, the idea of “elegance” is foreign to him for the same reason his answer was unburdened by chaos theory – he knows with precision, writes with precision, and understands the infinitely precise effects of his actions. Both walk away confused by the other, unable to come to an agreement because Ged doesn’t see any inherent value in writing rules that Hum can understand, and in fact, may assert that NO rule could satisfy the conditions he wanted while still being understandable to a finite mind (“elegant”).

        So, if we saw a road on Ged’s screen, that would road be the “Benign B” of:

        A: It is possible for an omniscient, omnipotent being to construct a “rule 21” that satisfies arbitrary criteria (if, of course, said criteria are not self-contradictory, equivalent to winning the “Mountain Game”).

        B: The arbitrary criteria are not met.

        D: The being, while capable of satisfying the given criteria, did not.

        The conclusion I want to draw, then (to avoid this Difficult D), is that the criteria must be self-contradictory. This invalidates D, since no being can beat the Mountain Game. In the case of the dwarves, this is the conflict between a perfectly sterile removal of roads and elegance of rule 21.

        But I fail to find an equivalent to “elegance of rule 21” in the theological analog, because elegance (a human construct) is valued by Hum because his finite mind can process it. When we define God as all-knowing and all-powerful, this construct ceases to have meaning in that context, unless we assert that God has a specific affinity for systems that are elegant (to human minds). And in the theological sense, God (with infinite knowledge) would be aware that elegant systems necessitate, in this case, sin/evil; the same way Ged is aware that any “elegant” attempt by Hum will fail due to the massive (or infinite) complexity required to remove the road in a sterile way.

        Therefore, I see another “Difficult D” when we move to the theological sense:

        A: The being has a self-contradictory set of goals/criteria.

        B: The being is not bound to the will of another, i.e. the criteria are, themselves, inventions of the being

        Lead to:

        D: This being created a self-contradictory set of goals or criteria for itself

        In effect, this would mean that God decided to prize the internal elegance of His system above the reduction of sin, evil, or suffering, at least in some cases. (I don’t know why we would assume He prefers elegant rules – for all we know, He could have the opposite goal/criterion, and despise rulesets that are human-understandable).

        The equivalent would be Ged creating a much simpler ruleset, with these accursed “roads” sometimes allowed, to satisfy Hum’s desire for simplicity. Ged, even with infinite capabilities, is bound by the external, uncontrollable rule of “Hum-understandability”, and that’s the reason he can’t create an elegant ruleset to produce fig. 4.

        But there’s no equivalent of Hum in the theological analog – and thus, proposition B holds for God. The conclusion, therefore, is that God himself is a conflicted being, unwilling to give up simplicity of rules (a goal we, again, don’t have any reason to believe He actually has) to prevent sin (something we DO have reason to believe he wants to prevent). I don’t know whether the biblical image of God contradicts this, but it doesn’t seem to be the image portrayed by the church I grew up in, and I have yet to meet a Christian who openly accepts what I think is the “Difficult D” of this line of reasoning.

        So, I think it’s the distinction between a “drastic” ruleset change and a “slight” ruleset change has me tripped up here. It makes sense with human analogies, but I think once you extend it to God, or any omniscient/omnipotent being, the inherent meaning is lost.

        My apologies for the long post (this probably could have been summed up much better in many fewer words), but I tried to cover all the pitfalls I’ve run into when discussing this before. And, I couldn’t help putting my thoughts in a story after watching your Memetics series 🙂

        I’m extremely interested in hearing your thoughts on this!

      • stanrock says :

        Great comment!

        Conceded: It’s true that Hum may prefer systemic simplicity and Ged may not.

        Not conceded: It doesn’t necessarily follow that an entity with no difficulty contemplating any arbitrarily-complicated system would not understand simplicity vs. complexity (any more than they wouldn’t understand that 2 is greater than 1), and/or would not prefer simplicity vs. complexity for aesthetic reasons.

        The article “A Theological Payoff of Meaning’s Funny Foundation” touches on this — it argues that “ultimately rational interests” is a contradiction in terms, so if we assume a Ged with interests, we have to assume axial/core interests that don’t necessarily have rationales (whether an interest in simplicity or a disinterest in creaturely suffering).

        (I’m not 100% proud of “Payoff” and don’t share it very often; you may have to meet me halfway.)


        Now, you mentioned that we have little reason to think simplicity would be one of these axial interests. In “Payoff” I proposed that nonintervention could be an axial interest. Here, I proposed systemic elegance or simplicity.

        In a funny way, these (elegant rules, and nonintervention) are two sides of the same coin. We can use the term “trap rule” to mean a fundamental system rule that is inelegantly tailored to detecting and reacting to extremely specific conditions, like a pre-programmed surgical robot ready to spring, imposing nonlocal exceptions to the natural procession of events. This functions just like a miracle, and because of its unpredictable (by creatures) functionality, would be supernatural for all intents and purposes.

        These days I think there’s a better description of an axial preference proposal, from which a disinterest in inelegant rules and/or intervention would derive. And that is the “teleological freedom of all creation, such that meaningful ‘others’ emerge, with whom to form relationships on the far end,” offered by deterministic chaos. Chaos acts as a kind of “math trick” that allows a wholly sovereign overseer to “let creation drift,” because chaos has an entropic effect upon detailed purpose information. So a creation that smashes against itself produces driftier, freer things as a function of that interference, in a compatibilistic sense of freedom that includes everything from ants, to asteroids, to attorneys.

        In other words, even though it’s all “dominoes,” the function of deterministic chaos takes a chainsaw to the “puppet strings,” naturally. Of course there are “big strings” that never get snipped, but the “little strings” get severed, too delicate for the violent and innumerable system collisions. Puppet strings would be re-tied and re-established as needed, but every imposition would make creation, effectively, “less free,” and the beings emergent of that creation less “teleologically-other.”

        This proposal has two features that take some “work” to accept:

        (1) It DOES involve an conflicted God having interest incommensurability within a manifold interest set. But you do see conflicted statements in Scripture, e.g., Ezekiel 33:11. In fact, all such statements of regret and lamentation can be expressed through this core conflict.

        (2) It makes freedom a subservient interest — freedom as a way to let folks drift, so that they are “other enough” such that genuine relationship-seeking is meaningful. (“Genuine” here is a subjective referent, and I’ve criticized the term when used in an “objective litmus” sense. It indeed comes down to interests that cannot have “ultimate rationalization.”) The Prodigal Son is let out of the house.

  2. Brent P says :

    > I will lose surgical control over that system, even if I’m 100% sovereign over it.

    But doesn’t your point about being able to erase the road highlight this is incorrect? You still have the surgical control to simply ‘erase the road’ but rather opt instead to be comfortable with the minor imperfections and flaws.

    How’s that different from the typical Calvanistic response that God’s Will must be content with the holocaust?

    • stanrock says :

      Your post rightly brings up something crucial.

      We have the premises that God exists, the Holocaust happened, and it wasn’t a power problem (e.g., it’s not as if Hitler overpowered God). As such it must in some sense appeal to the interests of God, but we have to be careful about loading our descriptions of this relationship with value statements we don’t intend (and load them with value statements we DO intend).

      Less of: Comfortable with, content with, the use of “decree” in typical Calvinist literature.

      More of: Suffers, bears, endures, stomachs — and labors in self-agonizing subtlety toward the eschatological hope of final restoration.

      An earlier article about these issues: “Is God the Author of Evil? (Semantics of ‘Want/Will’)”

      (My theodicean work here and elsewhere has “hints” of Calvinism, but I’m not a Calvinist — I think typical Calvinism has a number of “bugs” such that they fail to capture vital Scriptural revelation about God’s character and plans, and fail to fully extricate libertarian free will from their theology.)

      • Brent P says :

        I guess I don’t follow on how strongly God can be suffering/enduring under this system, rather than the free will system where there’s an arbitrary self-limiting of power in play (in order to service free will).

        Under this system, while God is not directly the author of the road, its presence is accountable entirely to him. He either has the option to remove it, or some higher desire that limits him (in the same way as a free will scenario), “I shall not micromanage” for instance.

        But this road is the product of those 20 initial rules. So were he limited by a higher desire, he still has the option to add #21 – “though shalt not build roads” into the deterministic programing from the start.

        I can’t seem to find my way to understanding how, like the free will system, Holocaust can be an undesired activity. God’s initial ruleset for the universe and later inaction make it not simply a possible consequence, but intentionally permitted.

      • stanrock says :

        Before continuing, I’d like to thank you for responding and taking the time to discuss.

        My sense is that even with libertarian free will, what happens and “sticks” is superordinately accountable to him, because he can arbitrarily redirect, prevent, or functionally undo anything that results from creaturely decisions — and, in Scripture, occasionally does (that is, he doesn’t have a “hard rule” that says “I shall never redirect, prevent, or undo creaturely decisions and their results”).

        I say “superordinately accountable” rather than “entirely accountable” because there are meaningful teleological senses where we do not, for example, account to God the motivations of malice and gross self-gratification that prompt many sins. God was not the sinner who sold Joseph into slavery; Joseph’s brothers were the authors. But God is credited with superordinate accountability (even such that Joseph hyperbolically utters, “It was not you, brothers, who sent me”).

        This distinction remains meaningful and preserved with no libertarian free will, but instead deterministic chaos. God is superordinately accountable for everything (just like under a LFW paradigm) and yet not “entirely” accountable for many specific things that have novel teleological emergence.

        You’re right that God can add surgically-architected rules, but if such a rule is made so specific as to avoid butterfly fallout, then this is functionally equivalent to fiat intervention. I suspect that “make no roads” is not incisive enough; I’m disinterested in a road at tick 13,500, but a transient road pattern earlier may have been acted as a necessary cog in the otherwise pleasing plan. The more specific the goal, the more an initial rule tailored exclusively to that goal (with no blooming fallout) will look exactly like fiat intervention (erasure).

        If that functional equivalence is true, then if he is interested in natural freedom such that his interventions are rare, then this drives a pruned and elegant ruleset as well. Put another way, inelegant special rules is “to-may-to” and fiat intervention is “to-mah-to”; you could code all miracles as hyper-special-case initial rules; that may be how it works, for all we know.

        Additional note: What allows for novel teleological emergence under deterministic chaos is also properly called self-limitation: Reluctance to intervene / reluctance to overengineer. But this self-limitation is wholly represented by God’s interests.

      • Brent P says :

        As always, you give me plenty to mull over and consider. Appreciate the detailed response!

  3. James says :

    May I save some of your articles to my hard drive?

Trackbacks / Pingbacks

  1. Does Determinism Make Us Robots? | stan rock - March 1, 2017

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: