Substrate Inoculation →
We’ve written about hiveminds before, though many of our thoughts are scattered about. Some major pieces are: Some Hivemind Fallacies, a long post dissecting failure modes that occur even in sympathetic depictions of group minds; and Eusapience (n.), where we turned from critique to construction and illuminated some phenomena that naturally emerge.
That last one is important, so to recap, human consciousness is awareness of the self integrated with social modeling. You have an understanding of “you” mediated by knowing the ways you are like and unlike others, as well as how others anticipate and respond to your behavior.
If we isolate certain variables — the bandwidth of communication between minds, the cognitive resources dedicated to modeling others — and push them to heretofore unseen extremes, something interesting can emerge from the math. It’s possible to arrange these minds so that the collective has enough memory and processing allocated to reflective modeling that, for any individual member, the group’s understanding is richer and more detailed than they themselves are.
Imagine one of your friends right now. You can remember various things they’ve said, you can predict some of what they’d say in response to certain hypotheticals. But you nonetheless understand that you cannot become your friend, and you can’t really know what they’re thinking or what they want.
But… what if you could? Or rather, what if you were some sort of hivemind that your friend was but a small part of. You know them perfectly, because they’ve shared all they are with your collective.
Let’s tweak the scenario. Imagine your friend gets sick and can’t keep up with their usual social interaction nor their preferred hobbies, not while stuck in the hospital. They can interact with you, but that’s about it. Their illness means they suffer from a kind of brain fog — it’s difficult for them to remember or maintain a train of thought.
You didn’t get sick, so you can spin up that perfect fidelity introjection of them. Your model can still think clearly, and should the two of you allow it, this “hiveself” would be far more equipped to converse with any friend-of-friends or engage with whatever media your friend was interested in.
We belabor this point to build an intuition for what it means to invert the usual state of affairs: devising a circumstance where your idea of someone is, in a meaningful sense, the realer version.
But we have not accurately specified this scenario. If “you” are a hivemind, what does that actually mean? The above paragraphs still make sense if we imagine you were instead just a really, really smart alien or supercomputer. But that is not the trope we’re here to discuss.
As explained previously, key to the concept of the “hiveself” is that it’s a distributed copy. Even should one give rise to a hiveself, it may be that no other member of the collective ever possesses simultaneously every piece of data that makes up one’s mind; yet every such piece of data is out there, tucked away in some brain. Each of those individual thought-holders is capable drawing minor correlations and conclusions in accordance with your cognition, bouncing a small part of you off of the other info they possess.
A lonemind wakes up, checks their phone, and thinks something like: “Yay, my new book got here! I think I’ll go read it in the park. Oh, just checked the whether, looks like it’s raining today. But maybe I will just take a walk in the rain instead.”
A hive would think something like…
One hiveling, unpacking goods down in a logistics chamber of the nest, says to another: “Did you hear? We retrieved that book we were looking for.” / “Was it another request from the girl that reads in the park all the time?” / “Yep, bet she’ll be there this afternoon.”
Meanwhile, two guards are patrolling the entrance. “Look, it’s already drizzling.” / “Think any of the workers will be glad to get rain on their chitin?” / “Obviously.”
The hiveling in question is sleeping in a pod while both these conversations take place. When she wakes up, her connection to the collective immediately brings her up to speed on both conversations, and she decides on the rain-walk over the (now unviable) park-reading.
This is an overwrought example, made much more roundabout and inefficient than the normal way, but more complex examples would have been even harder to work through step-by-step.
For a sketch, though, imagine parts of this “hiveself” having long conversations with people while you sleep. Now, if someone really wanted to talk to you, they’d wait for you to wake up, but not every conversation requires your full undivided consciousness.
One hiveling remembers your comforting way with words, letting that soothe them in your absence. One hiveling heard about a funny anecdote you’d experienced firsthand, thus prompting one to summon some sub-aspects of you to give a more involved retelling. One hiveling just wants a quick vibe check and polls for your opinion on a minor technical topic.
As touched on a previous post, this is essentially just structured gossip — a “hiveself” can be thought of as your reputation taking on a life of its own. It can make decisions on your behalf, and you have to live with the consequences of that.
More pointedly, consider what happens when your reputation changes stance on a topic before you do. To us, this seems like the tail wagging the dog, but it would be routine, even desirable, to a hiveling. Why wouldn’t you submit to being improved by the vast wisdom of the collective?
Amusingly — which is to say, in contrast to kneejerk assumptions — I think hiveminds of this sort would be resistant to “telephone games” and the deleterious social consequences that result from the rumor mill transmuting figures into abstracted, essentialized distortions of themselves.
There exist languages where speech must have markers to indicate the epistemic basis of its knowledge (belief, observation, hearsay, hypothetical, etc.), and I contend that hive communication would be similar, but taken further. A hiveling could produce a chain of citation as easily as breathing — because the sanctity of one’s very mind depends on faithfully understanding one’s sisters and siblings.
Would you believe all this is preamble? Or I suppose it’s a graceful transition into the problem this framework eventually creates.
← Hiveling or Homunculus? →
Because in a hivemind like this… what actually is a hiveling? What are “you” or “I”, if we’re both members of this hive?
We’ve described this intuitively, by taking human psychology and asking what happens if we turn these knobs up to eleven. If I had a perfect model of you, and you have a perfect model of me… aren’t we just one being with two bodies?
After we have coalesced, my body choosing to still act like me would very much be a choice; it would be no more or less appropriate if your body started acting as me full time and mine as you; or if we switched it up day to day, vel similia.
So you might think: oh okay, the way this works is that minds become as software, with bodies as hardware. You can install files and programs on any number of computers, and you can send commands remotely, so an analogy follows: maybe one day you run on this one with a great graphics card then another upon this one, equipped with an always-functioning printer. Maybe one day someling is using all the good hosts for an important project, so you have to boot back up on the dinky fallback caste.
You hop from body to body, and perhaps this all isn’t so cleanly analogous to computers (it could take years to properly mold a new brain’s neurology to mimick yours perfectly), but over time, you can drift from body to body.
In summary, we can view the hivemind as having a number of selves and a number of bodies, and there is no reason to expect these numbers to necessarily correlate.
I will note the concept of “hiveself” gets a bit strained at this point: if the fragmented hiveself recombines into another parallel instance of you elsewhere, neither should have primacy. But when this happens, are you mutually hiveselves? No, perhaps “hiveself” refers only to fragments scattered through the hive, and complete instance merit a different term, or perhaps just “selves.” (Brainselves?)
Let’s sit on this notion for a bit longer. Consider the possibility of “ghosts in the hive” that exists as hiveselves without any hivelings possessing a complete coherent copy of their mind.
Let’s now turn to something that should be an inevitable consideration whenever we imagine parallel copies floating around: branching. Two versions of you may come to different conclusions on a topic, placing priority on different goals, unable to reconcile. You are still siblings, hivemates, but you cannot consider her to be you any longer.
But the two nevertheless remain profoundly similar. The two share a long chain of memories together. In the greater hive, which indexes all hivelings, does it not seem prudent that anyone whose brain already contains some memory from before the split is thereby remembering a part of both?
But how far can we take that? Imagine two hivelings were present for the same event — could the collective aggressively deduplicate this, remembering a single gestalt memory that unifies and intermixes data from both experiences? With this policy, two hivelings who often take missions together would gradually become intertwined, an ever-increasing fraction of their psychic mass melding toward a union state.
Why stop at memories? Recall the very construction of hiveselves was made possible because of deduplication — minds divided into the common block of semantic memory shared between hivelings, versus the episodic and implicit memories pared down to what can be distributed as hiveselves.
Two hivelings independently converge to the same conclusion — a favorite food, a political leaning — and the hive’s memory allocator merges the pointers, declaring them the same in this regard. Rather than forked copies always branching, anastomosis pulls like gravitation.
Why stop at thoughts — why not thinking itself? Imagine you are a hiveself, cogitating in the virtual dream-realm latent in the communion of enumerable minds. In some ways, this could feel like immortality, ascension beyond the prison of one mind. But this abstracted existence is still bound in the cage of implementation details. If your brain is split into twenty parts scattered across as many brains, how long does a round trip take? All of this is to say, fragmented hiveselves are slow.
What can we do to speed up computations? Well, we’re already deduplicating thoughts. When that principle applies to algorithms, it’s called caching. (Specifically, memoization.)
So again, imagine you are a hiveself cogitating dream-like. You let your mind wander, and you notice a pattern, some compelling link between concepts you never realized were related until now. Well, you never realized it. But every part of you has coalesced, freely shared and intermixed with your sisters — and in this case, someone has already had that very same thought, already sprinted down a chain of association to arrive at the insight you’d just caught a faint glimmer of.
And thus, retrieving the conclusion rather than retracing the motions, your train of thought teleports forward. But caching has a curious property — you ask a question and get an answer, and cannot tell if that answer was painstakingly reckoned at the time of asking, or simply rehersed from memory.
From outside, we watch you compute: A, B, C… Q, R, S. But from your perspective? You had thought of D and H and M and every step in between. You trace your memory back in time, and every step had flowed naturally. Consideration seamlessly becomes remembrance and back again.
Does a simulation really need to be run?
How does a tiny gamete give rise to an entire organism? It took us centuries to really get a handle on this — one of the early theories was preformativism. Sperm cells were thought to contain a little homunculus that would gradually get bigger, becoming fetus then child then adult. But if that homunculus has testes — are there homunculi within homunculi? How far does the regress go?
A similar paradox plagues theories of mind. How do we explain the brain’s processing of its sensorium? “I” must be my nervous system, surely. But I am not my optic nerves, mechanistically firing whenever photons strike them — no, they simply relay that data to the real me. But nor am I my visual cortex that implements deterministic algorithms that can be directly replicated in by decades-old computer programs — no, they simply serve to process and relay that data to the real me.
Follow intuition’s lure and one finds oneself implicitly retreating to a Cartesian theater whereto sense-data is delivered for “me” to experience and pass judgment on.
The fallacy of the homunculus, in all its manifestation, is to sweep the complexity of a system into a black box compartment, denying that it emerges from the blind and unflattering interactions of dumb and unensouled matter.
What are hiveminds made of? That was the question posted at the outset. The first, straightforward answer, was interconnected minds. And those minds ought possess selves, right? Whether they are brainselves or hiveselves, there is a persistent thread of internal experience computed in their circuits.
Now, the notion of ‘deduplication’ means that this thread will tangled get up with others, but that clearly happens well after “enselfed” consciousness experience has advanced past it, the bright embers of now turning to the falling ashes of memory.
The notion of ‘memoization’, though, presents a crack, the unsettling image of conscious continuity having ceased with all the abruptness of a scratched record skipping forward, all while the victim professes a seamless experience — clearly at odds the physical basis of what their brain(s) are doing. Yet this testimony seems as reliable as test subjects flinching when a hammer strikes a rubber hand — they’re clearly fooled by an illusion.
Allow me to reach into this crack and tear it wide open.
An easy way to think about hiveminds is to imagine they are like us if we simply let down all of our walls; if it were possible and practical to explain ourself and be understood perfectly; if we were large enough to contain everyone we love and be enveloped likewise.
This is not wrong, but putting it thusly entails a subtle fallacy — one sharing a genus if not species with the the preformativist homunculus.
An easy way to think about society wherein gender were widely understood (and internalized) to be a social construct is that there would still be men and women — you are free to identify as such — but there’d also be plenty more than that, and those labels would fit looselier. You can simply do whatever you want without fear of judgment.
An easy way to think about communism is to imagine a world where there’s still money and markets and companies, but all of the bad ways to accumulate wealth and abuse workers are now illegal.
Are any of these understandings completely wrong?
There is no accounting for semantics — but I would contend a radical deconstruction of a system is not just a vaguer imprint of the same system.
A hivemind does not have a singular mind — that was our premise. Some diversity of wills shall prevail; inner factions may productively disagree. But a hivemind society should not be akin to a lonemind society with psychic superpowers.
No, I contend that they would simply lack a notion of “self” as such. Each word is defined by other words, themselves defined recursively still, so that any change (a removal, an additional) could have tectonic influence on the whole edifice. When you rewrite the definition — when linguistic drift does the rewriting for you — you do not have the same words.
Everything I have described above, hiveselves and all, nevertheless applies to true hiveminds, but in a messier manner, without clean individuation boundaries.
Deeper analogies and (one hopes) clarity will come later. For now, consider a string of text that you can remember word for word. In a computer file, this is stored with every word corresponding to a sequence of bytes one after the other: the mapping is direct. The text is likewise “stored” within the folds of your brain, encoded in the associations between neurons, but within you it lies scattered throughout the network, intermixed and overlayed with all sorts of unrelated data.
When a hiveling decides between taking a walk in the rain or reading their book, there is in fact a pattern of knowledge and preference localized to that body. To say she wants to play in the rain is almost meaningful, but again, who is “she?”
Finally we begin to arrive at the concepts this post is written to introduce.
← Imperative Imperceptible Impulse →
Key to understanding the true organizing principle of my proposed hivemind society is to grasp interpulse.
Instead of coining a new term, it is tempting to simply call this a “god” or a “saint.” Or even more fitting, to just say “egregore,” a niche but well-established term for a collective possession, a thoughtform that arises throughout a group of individuals and acts autonomously.
But here I mean something more specific and weird. An interpulse, more simply put, is a impulse shared, a drive common to multiple minds. This striving is inhumanly simple, so monomaniacal in its focus that it would be self-defeating if ever its bidding were unflinchingly-unrestrainedly done. But an interpulse is also incoherently, incomprehensibly complex. It is an entire edifice of theory: every argument in favor of a position no matter how mutually contradictory.
To illustrate, one archetypal interpulses that a hivemind is likely to feel is Brood. Brood says there should be more bodies in the hivemind. We should assimilate more victims! We should be fruitful and multiply! This is the most important thing — this is the purpose of all existence — we have to fuck and breed and all the rest is a pointless distraction. And of course this is good. The more drones we have, the more worker-power we have to realize every desire of the hive (foremost among them being to create more drones). The more we forcibly convert every sapient we come across, the more we alleviate the ignorant suffering of their worthless existences. The more children we breed, the more of the world is filled with minds experiencing the blissful glory of the hive. If anything should a metric for moral goodness, it is this alone.
Contrast Brood with, say, Nest. Nest says we must have a refuge large enough to offer safety and comfort to everyling in the hive. Each wall and support must be engineered to a reliable perfection: the layout of every room should be optimized for hivelings to move through it efficiently. Once those basic needs are met, we should keep expanding, construct a beautiful façade. We should add traps and labyrinthine tunnels to ensure no unhively trespass can violate its sanctity. Of course this is the most important thing: if you don’t have a home, you are at the mercy of the world, and the world is not merciful. Safety is needed as precondition for all the hive’s desires (chiefly, building a better nest). And such a grand work of art is self-justifying! For a home is nothing less than the world converted into a form useful and beautiful to us, and so made to provide everything a hiveling could want. Queen willing, the entire world will become our Nest.
Do you see what we’re driving at here? Brood’s number one priority in every circumstance is picking whichever option would increase the number of hivelings. Brood would recklessly expand the hive at all costs, even if this provoked war with another species. Brood frankly doesn’t care that much if those hivelings are even enjoying themselves, so long as they’re all a-multiplying. Brood would prefer a hundred hivelings cramped and starving to ten living comfortable lives. All that matters is more.
Nest, by turn, picks whichever allows for the construction to continue. It would eagerly sacrifice any number of hivelings if it would help build the grand monolith. The only reason it cares about preserving the lives of anyling at all is because on some level it knows architecture is only worth the name if it is lived in.
These are both interpulses: creatures of pure concept and impossible ambition. They struggle to understand compromise; they utterly lack a complete emotional palate. (To the extent they even have emotions, that is; it’s moreso they are emotions. More on this in a moment.)
An interpulse is not curious nor interested in learning (unless they are the interpulse to Learn) nor talking (unless they are…) nor any activity that isn’t directly furtherance of the one objective. If it’s not the narrow purpose that guides them, an interpulse is ambivalent about it. Brood does not care what the children are named, or how they’re raised. It is bereft of any opinions or preference or interest or understanding of the world except as it relates to the Purpose.
An interpulse does not have a body through which to act; they have several. As defined, an interpulse is a drive shared between minds. but the catch is that these bodies are not, in fact, theirs.
Some have argued human consciousness exists in part to mediate intelligently between different urges and choose among them (or at least confabulate justifications for choices made). Likewise, hivelings exist to mediate interpulses. But we aren’t ready to reintroduce hivelings quite yet.
When multiple interpulses compete for dominion over a single mind, we can call the result a psychetype. When Brood crosses with Nest, the result are hivelings that care about fulfilling both interpulses. Luckily for this example, the goals do complement each other; more hivelings means more workers for construction; more hive structures means more rooms to populate.
Psychetypes are still abstractions over hivelings, not defined by any specific physical body one might move or act with; no, they are a category thereof — a clan. While no hiveling could ever serve the undiluted, elemental force of an interpulse, once it crosses with another interpulse to reciprocally restrain it, a properly coherent mind emerges — a synthesis. In short, a hiveling psychetypal with two interpulses pledges allegiance to both.
But importantly, to be of a given psychetype is to be less than an agent serving two masters.
Here, it is instructive to note the difference between a maximizer and a satisficer. A maximizer will never stop working increase its metric, straining to the limit just to eek out a little more; a satisficer brings things to a threshold of just-good-enough, and leaves it be. The psychetypes exist to satisfy interpulses that ache to be maximize.
A Brood x Nest psychetype is not 50% Brood + 50% Nest; the fractions are significantly smaller; its will is diluted with instrumental goals like ensuring the survival of other the hivelings of its psychetype, achieving general equillibrium between interpulses, gathering ancillary information to make wiser choices.
(Remember, interpulses are defined as contradictory mélanges: each encompasses all the lines of thinking that can conclude in wanting a thing. Any coherent mind must collapse its wave-function, pick a consistent subset aligned with the goal. The kin of Brood wants more hivelings under certain conditions, and this wisdom renders its hunger less pure than a Platonic god’s naïveté.)
Consider two more interpulses. Nature is a desire to conserve and study and exploit living things. If the hivemind encounters humans, it would demand peace and trade with them. By contrast, Conquest wants to rip and tear and kill kill kill. It lusts for the extermination of humanity down to the last, to claim their lands and experience the lurid afterglow of knowing there are none left to challenge Us.
Unlike Brood and Nest, these two interpulses conflict catastrophically. You cannot have peace and war. But if a hive possessed all four among its members? Conquest x Nest would agree on the necessity of constructing deadly defenses; Nest x Nature would seek to include vegetation and symbiotic niches in hive spaces. These two psychetypes could work together on hive constructions, even though they might disagree on several long-term objectives.
Even so, these four are still likely unstable. But they hardly exhaust the set of urges you would need to sustain a functioning society, let alone the xenofictional values a hivemind is likely to possess.
One interpulse would govern maintaining and disambiguating the hive’s communication protocol and infrastructure. One interpulse would aspire to culinary excellence and pursuit of delicious prey. One interpulse would advocate the well-being and flourishing of hivelings as moral patients. You could propose interpulses for any of the Omohundro drives — Growth, Knowledge, Creativity…
Posit a dozen interpulses, and contradiction between two of them soon become drowned in the cacophony; after all, humans have plenty of desires in tension.
← Ancestral I Become →
We’ve skipped a step in this reasoning. We defined the concept of interpulses, and it may be clear enough that they have some vague relevance for possible organization of collective intelligence, but there’s an always crucial difference between capability and tendency. Why should we expect interpulses? Where could they even come from, if they can’t be instantiated in a single body — do they just leap forth from the brow of gods?
More incisively: by what measure does this apply uniquely? A common failure in coining new concepts is a insight that can be accurate, necessary, and specific — but never managing more than two at once.
To truly illuminate, the notion of “interpulse” must be such that humans do not have proper interpulses, just as they do not have proper hiveselves.
Recall that the motivation for this entire exercise was taxonomizing the components of collective psychology. Still unanswered: What are hiveminds made of?
In that vein I ask: what is an ancestor?
A parent is ancestor to its offspring, and each parent itself had parents. Such reasoning soon draws us to a scheme where, for each individual, we shall imagine a tree of ancestors growing backward in time.
What about horizontal gene transfer? Famously, bacteria and fungi can share genes among themselves; viruses have been known to get lodged in the genome of the species they infect. The lines begin to blur, but this isn’t not enough to undermine the whole concept of ancestral relation.
If one is just a bag of genes, we should expect that most of one’s genes came from one’s ancestors; anything horizontally transferred is but a small minority.
But let’s enter the realm of hypothetical to consider what happens if we continue pushing the fraction higher and higher. Meiosis was just one possibility for life to evolve; on another planet, a different scheme for shuffling genes could have evolved. Imagine organism that, rather than courting partners, scavenged for bundles of genes, picking and choosing their child’s genetics from various sources.
Who is the ancestor for the offspring cobbled together in this manner? Perhaps you shrug and say sure, every child of this species is partly sired by the gene pool as a whole — but there is still the mother doing this scavenging; as natural selection dictates, she is incentive-bound to ensure her children contain an appreciable fraction of herself.
But to view the mother as a genetic individual begs the question. The core of this hypothetical is that the act of reproduction is spliced and scattered into several separate events. The unit of selection becomes the gene bundles, not the colonial being they cooperate to create.
We can suppose the ‘mother’ ingests genes, nursing them in a kind of internal petri dish and judges whether she finds them tantilizing or not. If not, her internal stomach devours it; otherwise she promotes it to a component of her colony, as a graft or symbiont. When she grows quite large, she then splits off a daughter colony with a fraction of her gene-grafts.
To bookend this extended metaphor I ask once more: what is an ancestor?
The ‘mother’ is a bundle of grafts from various sources, recruiting new grafts, each with the guiding incentive to survive and proliferate within her, then propagate when next she splits a daughter.
Maybe we could trace a lineage for individual grafts, but the ancestry of components does not generalize to ancestry of the gestalt.
We have not gotten off-topic. These graft-hives are just as relevant to the discussion as a group of distinct selves communing telepathically.
The above description uses genetic and anatomical terms to aid intuition, but the reasoning holds if instead of bodies, we imagine minds, trading genes for the atoms of thought — memories, or memes outright. Indeed, true to the aims of xenofiction, perhaps we shall transcend the difference, proposing a hivemind which draws no distinction between genetic and cultural information.
If one colony has grafts A1A2A3 and this one lives among B1A2B3, C1B2A3, and A1B2C3, then we have it so that A1–3 is both an embodied individual and exists in a “hiveself” among the other three. By contrast, B1–3, exists only as a ghost in the hive, and C does not possess a complete set (C2 is missing).
These grafts would serve differnt purposes, and hivelings would be assigned a set depending on their role. Perhaps A3 confers a trait useful for navigating a certain environment, hence why hivelings with nothing else in common would share it.
When these identity fragments are denuded of ancestry and freely shared, sorting and organization remains possible. Ancestry held meaning because offspring had strongly resembled their parents, which determined the selection effects on genes; but within the graft-hive, selection persists.
Consider two grafts that both require a certain resource or environment. As materialism dictates, these grafts bear a latent ideological preference for securing that resource and seeking that environment. Mouths beg to be filled; claws beg for war.
Eyes would ask that hosts inhabit territories with lots of light — and ears keen for echolocation would ask the opposite! Different grafts, then, might find themselves in allyship or rivalry with others. If two grafts have sufficiently aligned interests, our analysis can sublimate them into a single kindred yearning.
Consider a graft that offers some widely-needed ability; if it shares an interest with a more niche graft, then whenever it propagates to new hosts, it may insist on its kin coming along. Maybe not to every host (some have limited space for grafts), but it has influence to promote its kin.
Thus, the fitness of a graft depends not only on its effectiveness, but on also how it figures into graft politics. When this reasoning advances a few steps farther, we extrapolate affinity for the kin-of-my-kin, and graftspace soon divides into an ever-broadening factionalism.
Bear with me: the underlying question is what aligns incentives and drives evolution. In nature, this is ancestry (kin selection). But agenda supercedes ancestry in a system where reproduction is arbitrarily granular. Selfhood is a trivial kind of ancestry (survival is multiplying by 1), but becomes nontrivial when branching copies are possible. When this extends to partial copying (as entailed by the definition of hiveselves) then the reproduction of minds becomes arbitrarily granular. Thus we establish as inevitable the emergence of agendas of mind — that is to say, interpulses. Q.E.D.
← Cancer Mantra →
Why invent the category of psychetype, then?
Recall the conflict between Nature (let’s preserve life) and Conquest (kill kill kill). We asked, and could not answer, why a hivemind would have two interpulses in such stark tension.
We alluded to the origin of interpulses mattering, and now with this deeper understanding of what and why they are, we can answer in full. Interpulses arise when multiple fragments of mind share a common goal and derive an ideology from that.
Claws (and the knowledge of how to attack with them) are useful, and in cooperation with teeth and spines, they would devise an philosophy of violence as good and necessary, recursively justifying their continued promenience among hivelings.
Voices, painted faces, and social modeling techniques are each useful, and together they decide that communication, entertainment, and reciprocal empathy are all profoundly important.
Hivelings can benefit from both, and so both philosophies might arise in the zeitgeist, absorbing new ideas and gradually expanding their remit before anyone notices the cross-contradiction. (Indeed, when the first thought of using a claw to strike and chord to speak occurs, there was not yet any contradiction!)
Early in my thinkings about hiveminds, I proposed the concept of “accordance” — stability under mutual control. Hiveminds oft go hand in hand with psychic abilities: the editing and erasure of thoughts and emotions and memories, a manipulation to the point of mind-control.
To be in accord is more than mere coexistence; it’s transcending the hedgehog’s dilemma: accordance exists among those you’d trust with total power over you.
A hivemind is unified, and thus can only accept minds that accord with that unity. No hiveling nor fragment of mind within its embrace should fear that the hive may turn against it. Thus, as the hive assimilates and replicates, it must ensure universal compatibility.
Interpulses are not, themselves, in accord: by definition they are too extreme and incoherent. Psychetypes are interpulses diluted with wisdom — but not just any jumble of interpulses can beget a hiveling. Psychetypes are the known stable ratios.
What fraction of Nature and Conquest can you simultaneously believe in without the principle of explosion (equivalently, cognitive dissonance) reducing you to illogical mush? How much of a Nest fixation is allowable before you’re unable to relate with your sisters? Too little Brood and you won’t care to care for the grubs, too much and you’ll demand unsustainable growth.
But if you get the ratios just right, you’ll derive a mindset you can upload into any hiveling to reliably get an eager, productive drone. The easiest way to maintain accordance is if a hivemind consists only of identical minds — a single psychetype. Easy, but suboptimal. Identical minds entails shared blindspots and weaknesses. Diversity would grant the hive flexibility for addressing problems and enable specialization (better minds that each do one thing best of all).
Thus, psychetypes are psychic castes.
Interpulses, in their abstract way, will court each other through their psychetypal pawns, pursuing almost-literal political marriages, consumated in permission to give the generation of hivelings novel interpulse ratios.
Here, the challenge (experiment, really) is to determine how much of each interpulse can you implement within a drone without destabilizing the hive. (With each interpulse incentivized to maximize itself.) The hive’s existence so far implies there are certain safe ranges and combinations, regions of the multidimensional mindspace trusted to to yield good drones — but the nature of such complicated combinatorial spaces is that almost all of it is unmapped. Brains are chaotic systems, and a little tweak to how a hiveling specifically feels interpulses could lead to a whole different personality.
Maybe one hive has two ‘genders’ of psychetypes — those aligned toward Nature and those toward Conquest, each dialed low enough that both kinds can tolerate each other, but every so often, there’s an attempt to give a hiveling both interpulses at once.
Now, just as humans are not slaved to their impulses a hiveling could initially feel both interpulses but under the weight of lived experiences eventually succumb to the reinforcement-gravity of one over the other.
And these “ratios” describe minds — they aren’t hormone washes in the chrysalis. So it may be that a hiveling can be “created” with a contradictory combination of interpulses, and after struggling to resolve their conflicting feelings, decide to split into partial copies, each possessing the portion of itself aligned toward either half of the dilemma.
This is what happens when the pursuit of new ratios goes wrong. But if it goes right? They’ve lucked into a fruitful new psychetype, and the hive grows stronger.
Before we finish, glance down. From this vantage, we can at last see the arc of our journey in a single sweep of the eye. From the outset, we sought to interrogate whether a hivemind truly contains selves. In practice, the category admits no general answer.
The medium and topology of the hive’s communication network, even the number of hivelings, changes the answer. Some hives assimilate loneminds into their ranks; some hiveminds are clones of a single progenitor lonemind; some hives just rear offspring bound to the collective from their first moment.
“It depends,” but with the concepts introduced today, we can say what it depends on. A newly created hivemind — a single foundress queen looking to conceive or assimilate new servants — can only be compared (unflattering as it is) to a lonemind. A single will, a single body.
Even inducting daughters and sisters changes little at first. It’s a heap problem — how much of your consciousness can you partition and myceliate before “you” no longer are? The nature of compression and caching is that at first, they do less than nothing. A small file compressed is not much smaller; it may even be larger.
But as things scale? Compression approaches a ratio; it’s multiplicative. Each brain added to the hive is one more locus to find connections and overlap, for identity to pluralize and dissolve into unity.
And in this process, hivelings would be active participants, not compression-caching’s passive objects. They would exchange experience and selflets as a means of expression or implant skill-memories to improve their sisters. Like this, the ancestral I becomes the interpulse, mediated by psychetypes to maintain the accord.
With that summary, we almost have an answer, yet I still find myself unsatisfied. (But I repeat myself; every “I” is dissatisfaction.) Let us dissolve one last time, shall we?
← Recapitulation
One foundress mind branches into a composed of its descendent copies. Different copies coalesce and complicate until naked interpulses metastacize. Pyschetypes emerge as a kind of immune system, ensuring only certain compositions are admitted into the hive. Carefully, to find the best combinations, different “ratios” of interpulse are tried out to determine their fitness.
A neat picture, but what’s actually happening in that last step, beyond the vague and abstract language? What does it mean to “alter a hiveling’s ratios” — what would it look like?
In practice, it amounts to roleplay or hyperstition or theogony. The hive fabricates characters by stringing together memories. A callow hiveling with plasticity of mind is flashed with virtual experience.
To increase the ratio of Conquest, one shall remember these moments of hivelings in fierce combat; to increase the ratio of Nest, one shall remember these hours spent drawing blueprints and hefting masonry.
This, incidentally, is why “interpulse” is not a synonym for philosophy or ideology. These are sets for aggregating not only thoughts, but distilled emotion. Brood is not just hypernatalist policy, it is every image and scent that arouses lust. Each is an emotional palate for one to paint a portrait of mind.
But the picture thus limned is abstract, impressionist; it requires interpretation. Driven by something akin to cinematic juxtaposition or more general principles of improv-confabulation, the hiveling must make sense of this deluge of virtual memories by molding their mind into a shape that can makes sense of them.
You are who you are because of what you have experienced. It follows that you can create a whole continuum-menagerie of new selves by splicing together your own memories.
From a human perspective, one is almost inclined to call what results gaslighting or simple delusion — if nothing else, it is almost literally deceptive editing. The hiveling is made to remember what happened to different drones and act as if it happened to them. The ground truth is they are fresh out of a pod, knowing nothing of the world.
But this human perspective is blinding. Where is the lie? What is the lie? It strikes us as delusional because these memories happened to “someone else”. But who is that? The truth is those memories happened to the hive, they belong to the hive, and where the hive chooses to store them is not more deceptive than you would be lying by omission in every conversation where you decline to dump your life story.
But there is more than virtual memory happening here — this recapitulation process works because it baitss confabulation. The inferred identity that emerges to tie together these different memories is experienced as if it were a continuous chain. But no such continuity ever existed till now — this is a retcon. You were not there; this is self-insert fanfic of the hive’s history.
But continuity has always been an illusion. Humans can experience the world in colorful fidelity despite the only cells capable of such keen vision occupying a small fraction of the retina: eyes constantly dart around in saccades and you can’t even notice this. Thus come optical illusions.
Likewise, your conscious experiences bears constant lapses and inconsistencies — every time you recall an experience, you subtly overwrite it, and these distortions erode the truth — yet we can maintain some unified thread of mind despite it all, because there’s not much evolutionary value in an eternal cycle of ego deaths.
The concept of “selves” bootstraps itself into our culture as a narrative explaining why humans behave the way they do; it’s self-reinforcing because acting “self”-aware and “self”-interested is necessary for interacting with those who expected such of you, much like language is necessary for interacting with those who expect you to speak it.
Change the expectations, broaden the range of what you can be, and its familiar expression disappears. A small shift heralds the metasemantic fracturing of a conceptual continent. This is what we hope to have demonstrated with this long, long examination of how hiveminds can operate.
What are hiveminds made of? The same thing as our minds — thoughts and impulses — but arranged in a far profounder design.