Unflattening Hobbes

In political science, the idea of a Hobbesian state of nature, featuring an endemic war of all against all, is a notional initial condition from which civilization could plausibly emerge. A generous reading of the model is that it is not about evolutionary realism, but about the plausibility of a pristine peaceful order emerging from a primordial violent chaos, under unfavorable assumptions about human nature (selfish and innately violent). In the classical Hobbesian model, the layers of the civilizational stack are bootstrapped from conditions that constitute a “flat world” in a social sense. Peace and structure evolve in parallel from this violently chaotic flatness.

But consider a conceptual alternative to the traditional Hobbesian model: what happens when we discard the assumption that structural order and endemic conflict are mutually exclusive? Or that peace goes with order and violence with chaos? Do we necessarily run into a contradiction? Could order emerge from chaos and endure, without peace necessarily emerging from war and enduring in parallel?

What if a Hobbesian condition of endemic war of all against all does not require the world to be a materially devastated and socially flat one, populated by warring packs led by grim young men in Henleys? What if it just feels like today’s world, but gets steadily slightly worse, slouching towards dystopia without ever arriving or unraveling? A Hobbesian end to history rather than beginning?

Last week, in Pack Experience, I proposed a model for the societal stack consists of five levels: individual, pack, troop, tribe, and imagined community, and argued that the pack level, rather than the individual level, is foundational.

My model of the unflattened Hobbesian all-against-all conflict condition is what you get when the actors at any level are able to act against actors at any other level, without necessarily collapsing the stack itself into a state of primordial or post-apocalyptic flatness. Pack experience becomes primal, but the world does not end.

In my view above (for the mathematically inclined, basically a full-rank outer product of the societal stack with itself, yielding a sort of societal tensor), societal order structures violence without being disrupted by it.

A school shooting for instance is an individual versus tribe conflict pattern where the individual wins. Racist policing is a tribe versus troop pattern where the tribe wins. An extra-judicial killing, such as that of Jamal Khashoggi last week by what we can assume were the tribal loyalists of Mohammad bin Salam, is an individual versus tribe pattern where the tribe won. In all cases, the societal order remains undisturbed.

You may not agree with my specific categorizations, or my particular taxonomy of levels, but I hope the logic is clear. You basically take the model of the societal stack and plot its levels along both the rows and columns of the matrix. So our 5-level stack yields 25 categories of conflict patterns (if we distinguish winning and losing actors for every unlike pairing).

Is it plausible that highly lawful and ordered states are capable of, and indeed, actively converge towards, deeper conditions of enduring violence? I want to argue that not only is such a condition possible, but that we might in fact be headed there unless we work to explicitly engineer a sustainably renewable peace. Unlike in classical Hobbesian theory, peace does not naturally emerge from mere societal structure as a positive externality.

The picture above is a matrix view of an unflattened, non-classical war of all-against-all Hobbesian world. Between the two extremes of stylized and limited flat violence that we recognize as such today — one-on-one MMA fights and declared war between nation states — there is a veritable Starbucks menu of possible patterns of formally undeclared, but sustainable, structured, and unlimited violence. The reason this can happen is that war has far more dimensions than peace. Just because there are more dimensions of peace does not mean the possibilities for war are exhausted.

This is a view of what I call second-order full-stack conflict, where any level of the societal stack can act against any other level of its own or competing stack. For this kind of condition to exist and be structurally stable, a fairly advanced technological state (with advanced communication and remote action technologies), and a built environment of infrastructure capable of surviving long periods of sustained conflict, repeatedly healing and regenerating into the same violence-prone forms, are necessary.

Unfortunately, I think we may be there already, and are complacently sitting around thinking the situation must necessarily lurch into a terminal crisis (and be fixed by apocalyptic heroism) or naturally heal itself into peace.

Neither need happen.

War and Peace, Flat and Tall

The conflict model I want to consider is, in a sense, the evil twin of the utopias favored by pacifist anarchists, who imagine (for instance) a sort of chaotic peace in a relatively socially flat and highly disaggregated world, with few layers, a minimalist built environment, and relatively weak connectivity among relatively small and self-sufficient packs and troops.

The opposite vision is obvious a world marked by richly structured conflict in a socially tall world, with many levels of aggregation, a maximalist built environment, and rich connectivity among building blocks of all sizes and scales. Let’s put together the obvious 2×2, plotting conflict level versus structure:

Classical Hobbesian models posit a flat war evolving into a tall peace, modulo some low-level biological bundling of individuals into packs.

Pacifist anarchists (as well as various kinds of libertarians and communitarians) fantasize about varieties of flat peace emerging out of a collapsing unsustainable state of tall war.

I want to consider what we might call a sustained state of tall war. Specifically, tall Hobbesian war of all against all. This does not mean everybody is fighting everybody else all the time, at all levels of aggregation. It merely means there is a coherent structure of societal levels, but no meaningful distinctions between combatants and civilians, or persistent macro-scale states of war and peace. Continuous partial conflict.

Tall Hobbesian war of all against all is a condition that is enabled, but not contained by societal structures, and does not pose an existential threat to the societal structures that enable it. It is marked by a rich diversity of patterns of conflict, multiple levels of technological sophistication, and active theaters of conflict at all scales, creating a fractal landscape of intertwingled war and peace.

Ursula LeGuin’s contrasting pair of worlds in The Dispossessed, Anarres (anarchic, flat peace) and Urras (orderly, tall brink-of-Hobbesian-war) illustrate specific imaginings of simpler top-left and bottom-right states. Techno-utopian visions tend to the top-right condition. You can plot your favorite science fiction worlds on this diagram. Try Asimov’s various worlds and Iain M. Banks’ Culture world for some fun.

You will notice that modern fictional near-future dystopian sci-fi worlds, such as those of Hunger Games, In Time, or Elysium, usually feature more than one kind of polity in juxtaposition. Most often, a center-periphery condition with a tall peace surrounded by a flat war. This is not an accident. World-builders correctly recognize that complex structure and uncontained conflict can co-exist.

Where they often go wrong is in assuming there is necessarily some sort of clear boundary between the two worlds, or that one can be created and easily policed. What we are learning today is that complex structure and uncontained conflict can inter-penetrate to the point of co-extensivity. Believe it or not, the Hunger Games world might be too utopian in its imagined patterns of conflict containment.

The Structure of Tall Conflict

What does a world of tall conflict look like? In real terms, it looks like Syria today.

In conceptual terms, it looks like my matrix view above, where there is active conflict among all possible pairs of levels, each pairing featuring its own distinct set of patterns. A world that is all Nakatomi Space, with no distinction between war zones and civilian territories.

The view above is a second-order view, where levels of different types can interact in an active conflict pattern.

The yellow diagonal, featuring like-against-like patterns of conflict, is a first-order world where there are boundaries between levels that prevent spillover of violence across levels, such as a feudal world where barons fight each other but do not burn crops. A zeroth-order world is one where violence is contained to a single level. Most imagined flat-world conflict states are zeroth-order: all individuals against all individuals, or all tribes against all tribes.

While the societal stack does not necessarily collapse, under conditions of unflattened Hobbesian conflict, trust does tend to collapse down to pack level, this is why the pack experience is central to the postindustrial societal landscape. As we speak, we are engineering trustless mechanisms for sustaining higher levels of societal order.

Put another way, classical Hobbesian models are an eventually flat model of conflict-driven societal evolution, where even if you don’t start out flat, you end up there. This assumption provides the primary motivation for seeking peace: you have something, in fact everything, to lose.

The non-classical models I’m proposing suggest that conflict can stay non-flat indefinitely, featuring a different kind of all-against-all conflict — a bleak cyberpunk landscape where all levels are robust to very high levels of conflict and can indefinitely endure conflict with all other levels. This means peace isn’t a necessary consequence of having a lot to lose any more than war is a necessary consequence of having nothing to lose. Peace is something that must be designed and sought separately, if indeed it is possible.

Let’s make this idea a little more precise.

In the classical Hobbesian all-against-all condition, flatness comprises two qualities: homogeneity and symmetry. The all-versus-all conflict involves actors who are all of the same type, and of roughly the same size. A stronger statement is that they are (or turn into) a field of actors of the smallest possible type and size. These correspond to first and zeroth order conflict conditions.

In the non-classical Hobbesian condition, homogeneity yields to heterogeneity (unlike actors can conflict, penetrating layers of containment if necessary) and symmetry yields to asymmetry (actors of different sizes can conflict). The conflict condition is second order at a minimum. 

A good way to understand this condition of increased ontological richness is as a new state of nature.

A New State of Nature

An unflattened Hobbesian conflict is a state of generative violence, where the structure continues to evolve into greater complexity, but violence simply takes on ever newer and more diverse forms, reaching towards the ultimate state of nature: true evolutionary Darwinian conflict.

This outcome is not necessary, but it is not impossible. Why?

It takes a moment to see this, but the whole Hobbesian model rests on the robustness of a distinction between natural and artificial environments.

So long as we can claim that nature is subtle, resilient, and complex, while technology is crude, fragile, and simple, we can argue that the two kinds of environment will evolve differently in response to human actions. When that distinction vanishes, the classical Hobbesian model fails. This is what is happening today. Actions that can destroy civilization can generally destroy nature too. Actions that do not threaten nature do not threaten civilization either.

Ironically, it is the very robustness of advanced civilization that makes it more conflict prone beyond a point of late evolution.

A sufficiently advanced civilization, noted Karl Schroeder, is indistinguishable from nature. If that’s the case, then technological advancement can (and I believe, does), lead to a decoupling between artificial structure and conflict level, and simultaneously, an erasure of boundaries between the natural and artificial.

Here’s a way to understand this: civilization switches sides.

The evolved societal structure and built environment that we thought of as playing on the human “side” in a war with the forces of “nature” switches sides. Nature and the structures of civilization ally with each other, blend together, and present a wholly new kind of formidable survival environment. We thought the world needed to collapse before nature could reclaim it. Turns out, it just needed to get sophisticated enough. Ashes to ashes, dust to dust, evolved complexity to evolved complexity.

We may start the game of civilization, but nature finishes it, by absorbing the built environment once the impedance mismatch between the subtlety of nature and the crudeness of technology vanishes, as is happening now. This leads to a rewilding of homo sapiens into what I’ve been calling a cyber-paleo species, with everything that implies.

So we could say that a new state of nature is emerging. One in which biological and artificial realities combine into a intertwingled, stable, regenerative landscape which cannot guarantee peace for the smug creatures that created it.

The more advanced the technology, the more unflattened patterns of conflict can continue indefinitely without threatening the stack itself. With software eating the world, this becomes even easier, since continuous cyber-conflict can be conducted with low damage to physical built environments, but high damage to psyches and well-being. Here is my matrix of cyber-conflict examples.

Higher-order conflict is obviously possible, and already here. The killing of Jamal Khashoggi features, besides a dissident individual and a tribalized neo-patrimonial state, complicated interactions with two additional actors: the OG Turkish Deep State and the newly neopatrimonialized United States under the Trump regime (there was briefly even a rumor going around that Khashoggi’s Apple Watch transmitted a physiological record of his killing; this seems to be untrue, but the fact that such a coupling between low-level tribal barbarism and sophisticated technological platforms is even plausible tells us a lot about the dimensionality of conflict).

Cyberization makes higher-order conflict even easier. A swatting for example is a case of an individual pwning a SWAT team (pack derived from an imagined community) against another individual. In terms of the ultimate combatants, it looks like a zeroth-order conflict (individual versus individual). In terms of chain of causation, the conflict involves a second-order pattern (SWAT team versus individual) triggered by another second-order pattern (individual-imagined community infrastructure hacking).

Arbitrarily high-order compound conflict patterns are probably possible, but the second-order matrix is complex enough for purposes of illustration (I suspect 3-interacting-level conflict patterns are relatively rare, so you can represent most conflicts as a sequence of 2-level conflict patterns, ie a trace on the 2d matrix that jumps from cell to cell).

Sticks and Stones and Phones and Drones

It is not a pleasant hypothesis. Grim though classical Hobbesian models are, they are fundamentally optimistic. War leads to peace as a side effect of chaos leading to order. That is a world designed by a benevolent god. Unfortunately, the case for that hypothesis looks increasingly weak. The Hobbesian state looks increasingly like a stable end of history state rather than an unstable primordial state.

The comforting intuition that societal structures cannot withstand sustained conflict conditions runs very deep, and is supported by a great deal of historical evidence.

A (possibly apocryphal) quote usually attributed to Einstein captures this intuition: “I do not know with what weapons World War III will be fought, but World War IV will be fought with sticks and stones.”

The conditions of industrial postmodernity are arguably already falsifying this intuition. The idea that endemic conflict destroys civilization has turned into a civilizational Noble Lie designed to preserve the peace.

The problem is that future need not be like the past, and the evolutionary dynamics that got us here need not keep us here. The end of history need not be a peaceful terminal asymptote. We can slouch towards dystopia.

World War III, as we now know, was actually fought through nuclear MAD postures, cold war arms races, economic competition, the installation of entrenched structural inequities, and endless proxy wars in a bipolar world.

The conflict in Syria — the first theater of World War IV perhaps —  is demonstrating that sticks and stones do not necessarily preclude cellphones and drones or even nukes. What are we to make of this?

My conclusion is that the idea of primordial flatness of Hobbesian conflict must be abandoned. Unfortunately for  #BTFSTTG (Burn the Fucking Shit to the Ground — note the implied flattening in the phrase) dystopians who are hoping for a quick, decisive slide into sticks-and-stones (and then presumably a rebuilding into a flat utopia), it appears that the world can remain standing longer than you can keep it burning.

Burning does not necessarily induce flattening. This was illustrated very clearly by the 9/11 attacks: a tall symbol was flattened, but the deeper societal structure underwriting it remained standing tall, and quickly rebuilt the symbolic structure.

Flattening a built environment — either physical or sociological — is not as easy as it looks. There is something potentially worse than a burning world that does not burn to the ground, allowing a better one to emerge from the ashes. This is a burning world that keeps on burning forever, but never to the ground, offering an increasingly diverse menu of ways to kill and be killed.

World War IV is not a war without nuclear weapons, unfortunately. It is a war where pack-level street fights among civilians feature sticks and stones and phones and drones, but nuclear weapons can still exist behind besieged walls, protected by satellites, robotic turrets, and minefields.

What does this mean for the future? It means that our experience of peace to date has merely been an accident of history; brief unplanned reprieves from unflattened Hobbesian war. What we thought was a free gift that accompanied civilization remains an aspirational technology we have yet to actually invent.

The fossil peace dividend of the Hobbesian evolution process has been used up. If we want more peace, we’ll have to figure out how to make it in a renewable way. We are experiencing more than physical climate change. We are experiencing sociological climate change. We need more than a transition from fossil fuels to renewable energy. We need a transition from Hobbesian evolution to engineered renewable peace.

For starters, I’ll leave you with a blank template of my matrix, for you to play Unflattening bingo with next time you catch up on the news. Or you can use it to trace out complex conflict chains that interest you.

Get Ribbonfarm in your inbox

Get new post updates by email

New post updates are sent out once a week

About Venkatesh Rao

Venkat is the founder and editor-in-chief of ribbonfarm. Follow him on Twitter


  1. Tetragrammarian says

    “Just because there are more dimensions of peace does not mean the possibilities for war are exhausted.”

    It seems like peace and war are in the wrong order in this sentence.

    • Suppose your state space has an entropy of 10, and you increase the entropy of your peaceful state subset to 6 from 5, then you reduce the entropy of your “nonpeace” subset to 4, if your entropy is 5000, and your entropy of peaceful states goes from 5 to 6, you don’t make much of a change to the range of possible no peace states.

      What on earth that state space actually is is another question, but perhaps something like tech and population enabled social configurations. Might make sense to think about the contrasing rates of entropy production actually, as the state space gets bigger over time. Venkat’s point rests in the idea that even as we uncover more peaceful states, and exclude warlike ones, we either also uncover new warlike ones, or in that quote, leave a vast gulf of warlike states non-excluded.

  2. influenced by Nick Sousanis’ Unflattening?

  3. Marc Hamann says

    I have to confess I found this piece to be quite irritating.

    While your systematic scheme might be a useful tool to think about patterns of conflict between groups of various sizes, the meta-level conclusions you are drawing about the state of the world are missing a couple important factors.

    First, you would need to explore equally the patterns of cooperation between these same group sizes, and then to distinguish the motivations and circumstances that lead to one or the other.

    Without exploring why a particular conflict exists, and why cooperation is not possible for those causes, you simply end up with a generic “the world is going to hell in a hand-basket” statement, which every era in the history of humanity has made.

  4. Carlos Bueno says

    I’m not sure this is new. Conflict is very useful to managing a society and maintaining control. A company keeps the salesfolk in a state of hungry mistrust to boost sales but also to erode their collective bargaining power. A kleptocrat keep two tribes at each other’s throats and so off his own. It’s intentional and maintained with layers of BS. Your yellow squares form the phase change line of Poe’s Law.

    The idea that conflict structurally harms society has never been true. Unsanctioned conflict, unmanaged hatreds, yes. The 70s was one long terror bombfest. Most of the perps were directed or sheltered in some way or another because they were useful.

  5. Good job Venkat!
    Whether this topic is new or not (and I think I do agree with Carlos mostly) I think you have said something useful.

    I am not sure I fully understand Marc’s point, but I do think our world is going to hell in a handbasket, so maybe that colors it for me.

    What caught my eye though, was that you seem to think that WWIII is over, and we are starting in on WWIV. This strikes me as much too optimistic. I think that the cold war and its echoes were about as peace-like as we can expect these days, and that WWIII has only just started. I first started thinking that when the Turks shot down that Russian fighter, but I see now that it is much more complicated than that.
    In any case, WWIII is still being fought mostly in world financial transactions so far, and hasn’t spilled into flying bullets much yet.
    I expect that to change really soon in the slower-than-we-think-possible trainwreck of history, but I wonder if you can tell us what made you decide that WWIII ended and that we’d started on WWIV?

  6. This is pretty good, I think the key point is an interesting one; that increasing the robustness of society can be an absolute disaster from the point of view of conflict incentives; the more the effects of our conflict can be contained, the less anyone cares to intervene to stop it happening, and the more possiblity there is of treating a given conflict as a useful asset to be preserved.

    Already sociologists are having a great old time analysing the strategies and networks of culture wars, as unlike in other conflicts, the effect of reputation and allegiance is significant enough that doxing, naming and shaming, and other forms of information warfare actually increase rather than decrease the legibility as far as slow careful analysis is concerned, as most disinformation is focused tactically and emotionally, rather than at voiding or obscuring a permanent social record.

    Conflicts fought with weapons leave behind dust, conflicts fought with emotional DDOS leave behind mappable similarities of messages.

    The question is of course, whether such research bounties help with peace, in the sense of spam filters or early warning signs, or in measuring the effectiveness of psychological warfare, and improving its targeting.

    The end result of this is probably the perfection of the weaponised dogwhistle, communication designed to be disruptive to the mental health of the target but not registering as such with anyone else, so that conflicts occur within their own closed communication chanels. Probably also accompanied by standard provision of mental health and recovery aids, drugs councilimg etc. underfunded but ubiquitous, similar to the modern embedding of independent healthcare within physical conflict zones.

  7. John Verdon says

    One thing you table is leaving out is ‘capital punishment’ – something that Christopher Boehm has discussed in some detail in Moral Origins. This would be tribe-individual – In think.

  8. John Verdon says

    One thing you table is leaving out is ‘capital punishment’ – something that Christopher Boehm has discussed in some detail in Moral Origins. This would be tribe-individual – In think.

    Also your 2×2 doesn’t collapse the contradictions but actually embodies mutually exclusive categories rather then embodied paradoxes of peace-as-war and war-as-peace or the social self as ‘I am because We are’ – the more connected (positively in peace and negatively in war) the more individual I am – and the more dependent I have to become on others to compliment my uniqueness (e.g. the new loneliness emerges with hyper-individuality emerging from hyper-connectedness).

  9. John Verdon says

    One thing you table is leaving out is ‘capital punishment’ – something that Christopher Boehm has discussed in some detail in Moral Origins. This would be tribe-individual – In think.

    Also your 2×2 doesn’t collapse the contradictions but actually embodies mutually exclusive categories rather then embodied paradoxes of peace-as-war and war-as-peace or the social self as ‘I am because We are’ – the more connected (positively in peace and negatively in war) the more individual I am – and the more dependent I have to become on others to compliment my uniqueness (e.g. the new loneliness emerges with hyper-individuality emerging from hyper-connectedness).

    This all presupposes that the ‘playing field’ is about humans vs humans – but the next leven in ‘tall war’ maybe neo-human vs humans – and as with Neanderthals – the winner is a hybrid.