The ‘Dune Hypothesis’



Join 36.9K other subscribers

I am an avid consumer of science fiction and fantasy novels. The most interesting aspect of such fiction to me is how authors construct social structures within which their heroes operate. Whether this happens in some alternate world where magic is possible, or “a long time ago in a galaxy far, far away” the authors need to create an internally consistent social reality. Some novels are so bad at this, that I stop reading them in disgust. But a surprising number of authors do quite a good job, probably because they have read enough history to internalize general mechanisms underlying the functioning of historical societies (cliodynamics), even if these authors would be unable to explicitly formulate these rules.

One example is Dune by Frank Herbert. I don’t know whether Herbert read Ibn Khaldun, but much of his cliodynamics, especially the aspects dealing with Arrakis and Fremen, could come directly from Ibn Khaldun. As a result, he creates a highly believable world (well, this is a science fiction novel), both ecologically and sociologically. This must have been an important reason why this novel was so successful.

There is one law of historical dynamics that Herbert discusses explicitly. This general rule may be formulated as follows: Harsh environmental conditions create a selective regime under which only the best survive, producing cultures with tough and capable warriors. This is the reason why the Emperor recruits his best shock troops, the Sardaukar, from the prison planet Salusa Secundus. Only the Fremen, evolving under equally harsh conditions of Arrakis, can match the ferocity and fighting ability of the Sardaukar.

What is particularly interesting about this hypothesis is that it is explicitly evolutionary. Nevertheless, I believe it is wrong. The problem is that it focuses on individual fighting ability, which is much less important than collective fighting ability. To give a single historical example, an average Roman legionary would most likely lose in a single combat against an average Celtic warrior. A Roman legion, on the other hand, would easily defeat an equal number of Gauls. Cooperation, discipline, ability to work as a team, willingness to sacrifice for the common good (in short, asabiya of Ibn Khaldun) is what wins battles and wars, not ferocity of individual warriors.

The selective regime that breeds militarily capable cultures is not harsh physical environment, but living in a ‘tough neighborhood.’ In other words, it is between-group selection, not individual selection, that creates aggressive expansionist cultures.

This post was prompted by a recent discussion with colleagues about whether people living in poor environments (those capable of supporting lower population densities) are more likely to go to war. The logic here is that people living under such conditions have greater incentive for attacking neighbors, than people living in rich environments. I think that in predicting incidence of warfare, incentives are less important than capabilities. So people living in relatively poor environments that are also characterized by intense between-group selection (e.g., Ibn Khaldunian Bedouins) would be expected to be quite troublesome for their neighbors. On the other hand, people living in poor environments with weak between-group selection (e.g., boreal forests) should be relatively peaceful.

I would be interested in hearing alternative views. And, ultimately it would be interesting to test the ‘Dune Hypothesis’ empirically.

Notify of
Most Voted
Newest Oldest
Inline Feedbacks
View all comments
Robert Kadar

Dear Peter,

I enjoyed your review of the science fiction novel Dune. It is now featured on Thanks!


Harvey Whitehouse

The Herbert Hypothesis is wrong. But it sounds right. Perhaps this is because we can all think of examples of warrior groups evolving in harsh environments, like the violent gangs of the ghetto or prison, the hill tribes of Afghanistan, or the fierce indigenous peoples of the Amazon jungle and New Guinea. There seems to be some correspondence between poverty and violence, between tough conditions and bellicosity. Peter Turchin makes the interesting suggestion that harsh environments may produce harsh individuals but not necessarily the most ferocious warriors. Perhaps, but imagine a map of Belfast with the areas of highest unemployment and poverty coloured red and the middle class, affluent neighbourhoods coloured blue. The red areas would demarcate the breeding ground for a great variety of paramilitary groups armed with guns and hammers whereas the blue areas would be populated with baggy-trousered liberals armed only with rolled up copies of The Guardian. Although tough neighbourhoods in Belfast are the niche for warrior groups, it is not obvious that they spawn more violent individuals. In fact, during the heyday of The Troubles the poorer parts of Belfast had exceptionally low rates of ordinary urban crime. Why? Because the paramilitaries were highly effective at policing those communities, suppressing unorganized or spontaneous criminality. This is probably why Herbert sounds right. Harsh environments like Northern Ireland have often produced highly cohesive warring tribes. But Herbert was also wrong because many harsh environments do not produce such groups. A possible explanation is as follows. Warfare is an extreme collective action problem, involving high risks for participants and strong temptations to defect. The most effective warrior groups are those that sustain the highest levels of social cohesion internally, overcoming the collective action problem. Warrior groups are likely to evolve in environments where social cohesion must be strong if individuals are to survive. Not all harsh environments are like that. In some deprived inner city neighbourhoods where there are no well-organized groups, or those groups are hard to join, the best survival strategy might be ‘every man for himself’, a recipe for individual harshness but not necessarily for tribal warfare. But in a place like Northern Ireland where everybody is automatically assigned to one or the other of two main tribes (Protestant or Catholic) group cohesion is more or less a birthright. In tough neighbourhoods like the Falls and the Shankill, this cohesion is converted into outgroup hostility because of intense competition between groups packed tightly into the same ecological niche. Conflict in turn amplifies cohesion in self-amplifying feeback loop. The result is sectarian warfare. But note that it is not the harshness of the environment per se that caused this situation. Harsh environments are often ones in which survival depends on solving collective action problems. The resulting cohesion is like a cocked gun ready to go off when the group comes into competition with others. So there may be a loose correlation between harsh environments and tribal warfare. In that sense, Herbert may have been wrong, but not completely wrong.

Peter Turchin

Harvey, I agree with your analysis of the Belfast situation. The critical element is that each group has the ‘Us vs. Them’ mentality. So group cohesion increases in situations where there is an intense and persistent conflict with other groups. Incidentally, religion very frequently provides the symbolic markers used by groups to distinguish ‘Us vs. Them.’ I have more on this in my book (War and Peace and War, especially where I discuss ‘metaethnic frontiers’).

However, I am still sceptical about the ‘struggle against environment’ generating the cocked gun you talked about. I cannot think of examples where such struggle requires cooperation in very large groups. Additionally, given that population numbers o f humans, like other animals, increase to fill the carrying capacity, this Malthusian factor will eventually cause hardship for groups living in both poor and rich environments.

David Sirrine

I disagree with your simplification of Herbert’s theme — there’s more to it than “harsh environments breed tough people.” The Sardaukar of Salusa Secundus and Fremen of Arrakis have more in common than their harsh environment; their hierarchical, survival-of-the-fittest social structure also affects their overall “toughness.” The Fremen are further toughened by the antagonism of the majority that fears and despises them, intra-group social pressures as severe as a pack of carnivores, and conflict with “neighbors” such as the spice worms. Herbert’s genius was that he at least unconsciously took into account factors such as cliodynamics more than you give him credit for.

Peter Turchin

Most of the factors you discuss will not result in higher asabiya (group cohesion). Intra-group pressures actually degrade it. The Fremen do not really fight against the spice worms – they worship them and ride them. The general rule is that it is conflict against other human groups that increases asabiya. The interaction between the Fremen and city-dwellers of Arrakis has the right structure to increase cohesion, and in fact it is very similar to the tension between ‘civilization’ and ‘desert’ in the Ibn Khaldunian scheme. Using the historical Maghreb as a model for Arrakis, we would expect that Fremen asabiya was high because of conflict with city dwellers and conflict between Fremen groups (if any). It’s been a while since I read the book, and I don’t remember whether the Fremen had a lot of tribal conflict among themselves prior to the unification by Paul Muad’Dib.

Michael Hochberg

I’m not sure to what extent group selection affects the likelihood that a group will wage war on another. This presupposes that the group in question has experienced past conflicts and has either been culturally/behaviorally modified to better prepare itself for future conflicts, or that (if inter-group selection) the group has actually been eliminated by conflicts, leaving behind the most “fit” groups. In either case (what are effectively cultural ‘mutations’ to existing groups emerge through experience, or at group founding), if this is to be Darwinian, then traits linked to warfare are selected if groups with certain traits survive better and spread.

However, it may not be past warfare at all that conditions traits driving warfare attributes. There may be endogenous features of a group’s environment that “select” for characteristics that make it more likely to succeed (i.e., persist, grow, and produce new groups): this may mean that these selected characteristics drive the group to be more belligerent, or possibly more conciliatory.

So the questions are:
1. Does the tendency to wage war come from individuals, quite independent of group consensus?
2. To the extent that individuals lead groups into war, or that there is consensus leading to instigating war, is this molded by natural selection over the lifetime of the group/culture in question?
3. To what extent do features of the environment, including resource productivity, internal unrest, scalar stress, etc… affect belligerency towards other groups?

I suspect that in small, egalitarian groups, (1) severe constraint, (2) perceived opportunity, and (3) perceived threat from other groups are all important in the war equation. On the other hand, in large, highly stratified societies, the information transfer and therefore ‘the reality’ of these same three factors are under some level of control by governments and media. So individuals or sub-groups determine whether the group as a whole goes to war.

In sum, it would indeed be interesting to see to what extent Darwinian selection (short term or long term), environmental conditions, and individual vs group consensus influence the instigation of war.

Peter Turchin


One thing to note is that external threats work on two time scales. On a long, evolutionary time scale ‘tough neighborhoods’ create strong selective pressures for groups with norms and institutions that turn them into effective war machines. This is a process of cultural group selection that operates in diverse ways: from genocide and ethnocide (or culturcide) through milder mechanisms such as imitation of successful groups, voting with the feet, etc.

On a shorter time scale, appearance of a serious external threat triggers in most societies a defensive response, which causes them to suppress internal divisions and increase social cohesion. This is sometimes known as the Simmel-Coser principle.

So we need to think of two sets of mechanisms, operating at different time scales: (1) relatively slow cultural evolution of norms and institutions and (2) rapid ‘behavioral’ responses dealing with immediate threats (or opportunities).

Michael Hochberg

Good points Peter. Clearly however, data is sorely needed. Just to repeat what I feel are the important questions:

1. Does the tendency to wage war come from individuals, quite independent of group consensus?
2. To the extent that individuals lead groups into war, or that there is consensus leading to instigating war, is this molded by natural selection over the lifetime of the group/culture in question?
3. To what extent do features of the environment, including resource productivity, internal unrest, scalar stress, etc… affect belligerency towards other groups?

Harvey Whitehouse

Peter, I think the presence of competing groups is crucial for outgroup hostility but insufficient to explain the levels of parochial altruism observed in highly cohesive military groups. Several different strands of research in social psychology suggest that willingness to fight and die for the group results from the cultivation of relational ties through shared experience and not simply from the presence of an external threat. Our ESRC project on ‘Ritual, Community, and Conflict’ is currently running a host of experiments and surveys among revolutionary brigades in Libya and other war-torn regions to investigate these issues further so watch this space. ☺ But as things currently stand I think the evidence suggests on balance that warriors don’t put their lives on the line for a cause, or to hurt the enemy, or even to advance the interests of the large-scale groups with which they may identify – rather, they put their lives on the line for each other, for the ‘band of brothers’. Sometimes this kind of cohesion does indeed appear to be fostered by harsh environments, but more generally it results from shared experiences of surmounting ordeals together. My colleagues and I are investigating the hypothesis that traumatic initiation rituals (common in warlike tribes and in the elite units of modern armies) artificially produce similar results. Outgroups may feature in histories of shared suffering, e.g. in the form of oppression and (yes) warfare, but there are many other ways in which bands of brothers are formed. Once that kind of cohesion is present it is indeed like a cocked gun – but it only goes off when the group is threatened. Once that happens, cycles of violence can perpetuate these dynamics. As you rightly pointed out in your initial post, however, the ultimate arbiter on all these theoretical issues will be the evidence we can muster. Until we have really disambiguated the mechanisms responsible for group cohesion, through experiments, surveys, quantitative analysis of historical and archaeological evidence, and careful observation on the ground, we are like Herbert the novelist just telling stories (even if his are admittedly much better than mine).

Vania Smrkovski

I’m not sure I agree with you assessment of the “Dune Hypothesis”. It seems to me that, yes, hard conditions does indeed breed a hearty people. But the Fremen do not rise up until they are corralled by the very organized system of the Atreides family.

Herbert goes to considerable effort to show that Paul learns from the Fremen, but that the Fremen learn from Paul as well, and are only able to rise up under the well-organized system of the Atreides House and the Bene Geserit ways.

So in essence, and perhaps subconsciously still, Herbert actually argues your point that the harsh conditions of the Sardaukar prison planet and of Arrakis are not, in themselves, adquate.

Peter Turchin

Ibn Khaldun, again, is a good guide to understanding these dynamics. The way I read him, constant strife among desert tribes creates very cohesive and military capable groups. But their asabiya is somewhat narrow – it is great for building cohesion at the level of a tribe, but not so great for building highly cohesive tribal confederations. Here is where Ibn Klhadun brings in religion – it provides basis for cooperation at a much larger scale, at the supratribal level. The best example, of course, was Prophet Muhammad’s unification of Arabia.

So in Dune, Paul Atreides benefits from prophecies implanted by Bene Geserit in his quest to unify the Fremen, which triggers the jihad – in many ways similar to the historical explosion of Islam in the seventh century.

Juan Alfonso del Busto

I find this topic interesting too, though I am not so sure an individual celtic warrior could defeat a well trained roman legion soldier in a one to one fight…

Speaking of Science fiction, there is a short story by the great Isaac Asimov that brings out this debate. The story´s name is “In a good cause…” and the logic behind the plot goes more or less like this: before the roman legions ever existed, the hardest soldiers and mercenaries were the Greeks. In Asimov´s story a whole futuristic interplanetary human society is deliberatedly but unadvertedly driven by a visionary man called Geoffrey Stock towards a hard internal competition, a sort of civil war, in order to mold that society into a greek-like war machinery capable of defeating an alien enemy society: the diaboli. When the generalised war starts between humans and diaboli, the internally quarrelling human society discovers that it is in much better disposition to fight and win than the more armonious and internally cooperative alien society. Though Stock is the real savior of humanity, people will praise his friend and opposer Richard Altmayer for deffending the unity of humanity against the diaboli from the beginning.

Montesquieu wrote in “Causes of the geatness of the romans” that nations that just come out of a civil war are stronger, not weaker, than before.

How did the ancient Greeks become such good warriors that could overcome much larger barbarian (usually persian) armies (as Xenophon´s “Anabasis” proved and Alexander the great conquests confirmed)? the answer probably is a climate of extremely intense intergroup competition among state-cities, in the ecological niche called Greece, which selected increasingly better organizational configurations for battle… and government. The Greeks would have never be in shape for defeating the barbarians if they hadn´t been fighting themselves for centuries. The greek “phalanx”was effective mainly because of its internal organization and cooperative spirit (the theban sacred phalanx were composed of pairs of “lovers” who had sworn to fight to the death), but also because of the individual superior training and equipment. The greek hoplites were heavily armed and well trained because they could afford it. The hoplites came from the aristocratic classes (supported by the working classes and slaves who undoubtedly led the real harsh lives). Therefore general social structure is important too. A social structure that encourages aristocratic men to quarrel and achieve military glory through discipline, cooperation and bravery (and that socially punishes those who avoid these duties) in a intense intergroup competition context is likely to produce a nice phalanx-like battle configuration. This is specially true if population density is relatively low (in other words: human meat is expensive) and the orography of the intergroup “arena” facilitates cultural diversity and between group differentiation.

And there is sort of darwinism at the individual level too. Let´s not forget that spartan society fostered a kind of eugenesics: women were allowed to mate with men taller, stronger or braver than their husbands.

Of course, imitation between groups is also key to the success of the whole. War innovations in ancient Greece were quickly adopted by avid observers, what granted a high rate of innovations to spread.

This is also very interesting because it proves that internal competition is necessary to improve the whole. A perfectly internally cooperative organism can´t evolve new adaptations within its lifespan… unless it has built-in Plotkinian Darwin machines (like inmune systems and learning nervous systems). Internal “controlled” competition can be used to generate adaptations in a pre-existant organism within its lifespan. Otherwise, the only way an organism has to evolve adaptations is to multiply with variation and die, “hoping” that the new generations will have “what it takes” to cope with the new problems in the first place.

Peter Turchin

The effects of internal competition is a big topic that requires a separate blog to address. I’ll try to get to it soon.

On the putative military prowess of the Greeks. Remember that most of what we know about ancient Greeks comes from the writings … by the ancient Greeks themselves. Naturally, the Greeks thought that the way they did things was the best. It is especially funny that they considered highly civilized Persians ‘barbarians.’

As to how successful the Greeks were at fighting against the Persian empire, remember that by 500 BCE the Persians conquered half of the Hellenic World (most notably the Greek cities on the eastern shore of the Aegean). After that they also annexed Macedon. True, the Persians experienced certain difficulties with military operations against Athens and allies in the early fifth century BCE, but the main difficulty was logistics – Athens fought a few days of march from home, while Persians fought 3,000 km away from the empire’s center (Persepolis). Greek tales about invading Persian armies numbering in hundreds of thousands are simply not credible. In any case, by the end of the fifth century Sparta became essentially a client of Persia and, using the Persian subsidies, constructed a fleet which allowed the Spartans to win the Peloponessian war. So by 400 BCE all of the Hellenic world was within the Persian imperial sphere, in one way or another.

Juan Alfonso del Busto

Of course you are right. Ancient western history was written by the Greeks, who took good care of picturing themselves as the best. The persians also had good reasons to consider themselves the best; for instance they had the elite corps guard called “the immortals”, which surely matched the best Greek or madedonian hoplites phalanx (on the other hand, the immortals guard is another example of an aristocratic, professional, well trained and heavy equipped troop). And it is also important to remember that the original persians did come from a harsh environment, the Zagros mountains, and that they were able to conquer the medes, the lidians and the caldeans during “Cyrus the great” lifetime. Indeed, this seems to be a common thing in History: the hardest soldiers usually come from the mountains or similarly harsh environments, like the numids who fought for the carthaginians, the germanics who fought for the roman empire, the isaurians who fought for the bizantine empire, and so forth.

I am not a “Dune” follower, so my question is: Do the Fremen have a complex hierarchical social structure? Do Fremen warriors come from the aristocratic classes or they are just “the common men”? As far as I know the latter is not very usual in History. I think that only the slavics were kind of peasant-warriors.

However the milestone achieved by Xenophon´s 10.000 must be taken into account. They survived more than 2.000 miles away from Greece and came back home from that hostile land with no logistics whatsoever. But the most remarkable thing achieved by them was that they didn´t split at the beginning of their journey, when the persians treacherously murdered they leaders. The greek mercenaries quickly found another leader (allegedly Xenophon) who succesufully guided them back home. The persians didn´t expect that: they thought that the body couldn´t survive without the head. They were wrong. This is the magic of democracy in action (even though the athenian Xenophon himself was pro-spartan and pro-oligarchy). This is the reason why I say that social general structure and government institutions are important too. Within-group stability provided by democratic values, beliefs and ideas (Boyd and Richerson cultural variants or memes) created real long term cohesion… at least during their oddissey, while the 10.000 were surrounded by enemies… They dissolved as soon as they got home.

And we can´t forget that romans emulated greek culture (including the war tactics), and they were extremely succesful conquerors. But again, the war between Romans (and bizantines) and Partians was epic.

Gul Deniz Salali

Rituals are likely to be part of the picture but we need to be clear about where they fit in the causal chain. I think of them as a sort of ‘social technology’ that groups can use. Technologies, however, emerge when a need arises and as Peter already pointed out, this need arises for groups when they encounter a threat. Would there be Nationalists if there weren’t Unionists in Northern Ireland? Would there be highly cohesive military groups if there was not inequality or corruption in Arab world? I think the need for rituals or costly initiations comes after the emergence of a threat.

Returning back to our initial discussion point about whether people living in poor environments (i.e. environments with low carrying capacities) are more likely to go to war, I agree with the Peter’s point that people living in harsh physical environments and experiencing conflicts with external groups would be expected to be more competitive. However, I wonder what would be the odds of victory for those groups on harsh environments if they were to compete with the groups with better resources (e.g. the groups living on more productive lands).

Peter Turchin

Deniz, your point about the probability of victory when people from poor environments attack neighbors living in productive environments is very important. Generally speaking, groups from poor environments will have fewer resources for successful attack – lower populations, inferior weapons and armor, etc. Because people are not stupid, they will calculate that the likelihood of successful attack is low and that will deter them. So if they do attack, they are more likely to act as bandits than an invading army.

  1. Home
  2. /
  3. Cliodynamica
  4. /
  5. Regular Posts
  6. /
  7. The ‘Dune Hypothesis’

© Peter Turchin 2023 All rights reserved

Privacy Policy