This series is now available in audio format. You can find the playlist here.
This is the fourth part of a four part (I, II, III, IV) look at the Dothraki from George R. R. Martin’s A Song of Ice and Fire and HBO’s Game of Thrones. We’re looking at, in particular, if Martin’s claim that the Dothraki are “an amalgam of a number of steppe and plains cultures” can be sustained in the face of even basic knowledge about historical Steppe and Great Plains nomadic peoples.
Last week, we concluded that the vast majority of Dothraki culture, social organization, economic practices and family structure are effectively completely untethered from the historical realities of effectively any of the literally dozens of historical Great Plains Native Americans or Steppe nomads. This week, we’re going to close out our look by discussing Dothraki warfare. We’ll start with the visual – weapons and armor – and then move to the conceptual – strategy, operations and tactics.
And as always, if you like what you are reading here, please share it; if you really like it, you can support me on Patreon. And if you want updates whenever a new post appears, you can click below for email updates or follow me on twitter (@BretDevereaux) for updates as to new posts as well as my occasional ancient history, foreign policy or military history musings.
Finally, as a reminder both of what we are investigating, the key statement we are really assessing here is this one by George R.R. Martin:
The Dothraki were actually fashioned as an amalgam of a number of steppe and plains cultures… Mongols and Huns, certainly, but also Alans, Sioux, Cheyenne, and various other Amerindian tribes… seasoned with a dash of pure fantasy.
It is not the existence of a fantasy culture which draws our attention, but the explicit declaration that this fantasy culture is not merely inspired, but ‘fashioned as an amalgam’ of real cultures, which both existed in the past and still exist today, with only ‘a dash of pure fantasy.’ That line is important, to be clear, because it presents the fictional Dothraki as a statement on historical Native American and Eurasian nomads and – when combined with Martin’s statements that he relies on history to inform his work – that this statement is based in some sort of historical reality.
Which it isn’t. But we’re getting ahead of ourselves.
Where There’s a Whip…
The Dothraki are described as having three main weapons: bows (AGoT, 86, 555, 558, 597, 669), whips (AGoT, 86, 194, 493, 555, 596, 669) and a curved sword called an arakh (AGoT 85, 86, 327, 493, 555, 556, 559, 560, 596, 597, 669, 674); of these, the arakh is clearly the most prominent (I am sure I have missed a reference to a weapon here or there, but I hope the citations here give some sense of the relative weight each is given – the arakh is the most frequently mentioned by some distance). When a Dothraki warrior enters Vaes Dothrak, each, “unbelted his arakh and handed it to a waiting slave, and any other weapons he carried as well” – after the arakh, the other weapons are seemingly afterthoughts (AGoT, 327). The prominence of the arakh in the narrative is underscored by the fact that it is the only one of these weapons whose name we learn in Dothraki, or which is described in terms of its shape or special function (AGoT, 85), while the bows and whips remain just bows and whips (ironic, as it was Steppe bows, not Steppe swords, which were unusual).
We might dismiss this as simply an accident of Daenerys’ perspective – that, being Westerosi, she focuses on the weapon most meaningful to the Westerosi – but that’s clearly not true. After all, the offering of an arakh is how Daenerys’ loyal followers demonstrate their fealty to her, in a ceremony that is clearly Dothraki, not Westerosi (AGoT, 674). It is also, I should note, the only weapon we see non-Dothraki using that is clearly identified as being foreign and typical of the Dothraki. It remains special through the eyes of multiple point-of-view characters, including military men.
(And, as an aside, now that we are this far in, it seems obvious but worth saying that the fact that Martin has no Dothraki viewpoint characters in his narrative is hardly a saving grace; it merely intensifies the ‘view of a savage culture from outside’ effect. As we’ll see, this makes perfect sense given what seem to be the actual inspirations for his depiction.)
The prominence of a curved iron (or steel) sword lets us rule out a Great Plains Native American inspiration for this kit right out; the sword was never a significant part of Plains Native American armament (the lack of tool-metal production in the Americas prior to European contact means that there was no indigenous sword-making tradition, although the maquahuitl represents a clever sort of ‘sharpened club’ design). Even after contact, it’s hard to avoid the conclusion that the expense of trading for a sword wouldn’t have been justified by its utility over a steel axe which might also double as a tool (on axes, see W. Lee, “The Military Revolution of Native North America: Firearms, Forts and Politics” in Empires and Indigenes (2011), 62-3). So we must turn to the Eurasian Steppe.
And immediately we run into problems, not that any of these weapons are wrong per se, but that their proportion and prominence is all mixed up and that there are other, far more important weapons missing.
For a Steppe nomad, by far, above and away, the most important weapon was the bow. The Armenians literally called the Mongols “the nation of archers” (May, Mongol Art of War, 43). Nomads spent the most time learning the bow (May, op. cit. 42-49) and it was the one indispensable weapon. Indeed, so indispensable that nomads were generally required to have several; the Liao Shi records that Khitan nomad warriors were required to possess four bows and 400 arrows, while John de Plano Carpini reports that the Mongols all needed to have 2-3 bows and three larger quivers (May, op. cit. 49-50). The Steppe bow itself would also have looked unusual in both shape and construction to a Westerosi observer either strung or unstrung – they were composite bows, made with a wood core, a backing of horn and a rigid end-piece (called a siyah in Arabic) and were generally drawn with the use of a thumb-ring to reduce strain on the thumb (May, op. cit., 50-1). This unique construction allowed these bows to reach draw weights and launch energies equivalent to the far larger yew longbows of England and Wales and still be compact enough to use from horseback.
(I should note that the bow was also the paramount weapon for the Native American horse-borne nomads of the Great Plains, at least until it came into competition with firearms, though my understanding is that Native American bows were not as powerful as Steppe bows).
But even after the bow, the sword is not first. Or even close to first. Or, indeed, even on the list! The Khitan regulations I mentioned included four bows, two spears (one ‘long’ and one ‘short’), a club, an axe and a halberd, but no sword. John de Plano Carpini describes the full kit as two or three bows with quivers, an axe, ropes, and swords only for the wealthy (May. op. cit., 50). Speaking more broadly, May notes that spears (used as lances from horseback) seem universal in accounts of the Mongols, but “accounts are contradictory regarding whether these [swords] were universally used” (May, op. cit., 52). While May supposes that the ughurgh-a, the Mongolian lasso, might have been used in combat – and it may well have – we have no definitive evidence of it. If it was ever a weapon, it doesn’t seem to have been an important one.
In short, while the Dothraki’s weapons are an arakh-sword, a whip, and a bow in that order, the Mongol’s chief weapons were his bow, followed by his backup bow, followed by his other backup bow, followed by his spear, and then his axe and only then followed by a sword, should he have one, which he might well not. The reason for preferring an axe or a spear for the humble nomad should not be too surprising – iron in quantity could be hard to get on the Steppe. Spears and axes are not only weapons, but also useful hunting and survival tools; swords are generally weapons only. Nomads generally cannot do their own metal working, so swords would have to be imported. Moreover, even in a melee, the first recourse would be to a spear, whose reach on horseback was a huge advantage, making a sword an expensive imported foreign luxury backup weapon with no additional utility. Nevertheless, it’s clear that Steppe nomads, once successful and moving into agrarian areas, liked to acquire swords – swords are effective weapons! – but the sword was about the furthest thing from the core of Mongol culture the way the arakh is practically the symbol of Dothraki culture.
The other issue, of course, is the arakh itself. Martin describes the weapons as “long razor-sharp blades, half sword and half scythe” (AGoT, 85) and goes back to that scythe analogy (e.g. ASoS, 245). It seems generally asserted that what Martin means by this is something close to a scimitar (I have to confess, I haven’t found anywhere that Martin says this, but I’ll take the reader consensus). A scimitar of some sort (the term refers not to a specific form of sword, but a whole family of curved sabres, almost all originating in Asia) is the correct sword. Mongol swords were, John de Palno Carpini tells us, “pointed at the end but sharp only on one side and somewhat curved” (May. op. cit., 50), something like a Turkish kilij or a Persian shamshir (both forms of scimitar), though given his description, perhaps not as strongly curved as some of the examples of those types.
I have to admit, ‘scythe-sword’ (ASoS, 245) is a really strange way to describe most of the weapons in the scimitar ‘family’ (which includes a number of different curved sabres from Asia), though. A scythe-blade faces the wrong direction, but it is also sharp on the wrong side – scythes are sharp on the inside of the curve, whereas scimitars are sharp on the outside of the curve. There are swords with sharp edges on the inside of the curve (I tend to class these as ‘forward curving’ swords due to the direction of the curve when the sword is held), such as the Greek kopis, the Spanish falcata and the Nepalese kukri; of these, only the kopis seems to have been a cavalryman’s weapon (Xen. On Horsemanship 12.11). These forward curving weapons, being shorter and stockier, are clearly not what was intended by the arakh, which is consistently described as long (e.g. AGoT, 85, ADwD, 884).
Instead, the scythe metaphor fits the overall framing of the arakh, a weapon “better to cull the infantry ranks without breaking stride,” a “murderous blade against half-naked foes,” (ADwD, 884), a “wickedly sharp scythe-sword” (ASoS, 245), ineffective against armored opponents. Not an elegant, fine weapon, but a cruel ‘murderous’ one, made for ‘culling’ unarmored infantry and peasants, as one reaps wheat or hay. I don’t want to push this point too far – in all of these many pages, the arakh simply doesn’t get enough characterization to make the case watertight – but the characterization it does get all seems to push in this same direction: a murderous weapon for a murderous people…which of course fits with effectively all of the other characterization the Dothraki have been given. On the balance, I think Martin is a skilled enough writer to understand the implications of the scythe-sword description and to have intended them (and then subsequently reinforced them).
Nevertheless, credit where credit is due, while the place of the arakh is entirely out of all sensible proportion with how it would be considered by actual nomads, it is the correct sort of sword for a steppe nomad (if we assume it is, in fact, a scimitar of sorts). That said, prioritizing the arakh belies a fundamental misunderstanding of how Steppe (or Plains Native American, for that matter) warfare and culture worked. Placing the arakhat the front is thus indicative of deeper problems.
Of course we couldn’t leave off without discussing the absolutely bizarre visual adaptation of the weapon for Game of Thrones, where the scimitar-like arakh is transformed into what is essentially an oversized iron khopesh:
Presumably what happened here was that someone read ‘scythe-sword’ who had never seen a scythe, but had seen a sickle, and decided that a sickle-sword was the way to go, but also wasn’t aware that sickles are sharp on the inside of the curve and not on the outside of the curve, and so went with a forward-curving ‘sickle-sword’ design (which is sharp on the outside of the curve). And then, for good measure, inexplicably chose a short weapon made for the bronze age and just scaled it up to absolutely massive size. Moreover, the show’s version of the arakh inexplicably has a long, two-handed hilt, supremely impractical from horseback.
There are so many problems here. First, the khopesh is more of an axe-sword than a sickle-sword. Moreover its form is directly connected to the material properties of bronze. Because bronze doesn’t resist bending as well as iron, bronze swords need to be short and the khopesh was generally below – often well below – 60cm in total length (markedly smaller than, for instance, a gladius; do note that the gladius and the khopesh were never on the same battlefields, they are separated by almost 1,000 years). While the Ethiopian shotel is a tempting comparison point for an iron ‘sickle-sword,’ unlike the arakh (or a khopesh), the shotel is actually a sickle-sword, sharpened on the inside of its curve and used for hooking attacks; the arakh of the show is clearly not wielded like a shotel – instead attacks are mostly made with the outside of the curve (though both might be sharp?).
By any measure, the result is a terrible weapon. Weapon designs cannot simple be ‘scaled up’ like this without ruining the things that made them successful; a jumbo-sized khopesh is almost guaranteed to be too heavy. Unlike most scimitars – note John de Plano Carpini on the sharp points of Mongol swords above – it cannot give point, which (contra Xenophon) is a real disadvantage on horseback. It looks to be a two-handed weapon for use on horseback where one-handed weapons are most appropriate. As a two-handed weapon, it has inferior reach for what must be its considerable weight, and the forward curving shape offers it essentially no cutting advantage, unlike the kukri, kopis or falcata, which lean into their cuts.
I haven’t any idea why they opted to adapt the weapon this way, except to note that it fits in with the general pattern of the show taking Martin’s already cringe-worthy exoticism in treating the Dothraki and dialing it up to 11.
The Dothraki attitude towards armor is made fairly clear. “The Dothraki 1 had mocked him [Jorah] for a coward when he donned his armor” (AGoT, 556), something echoed later by the arakh-wielding but Merenese (that is, not Dothraki) Khrazz (ADwD, 885). The show extends this out to a discussion between Rakharo and Jorah (S1E3, 48:10ff) , that “Dothraki don’t wear steel dresses” (though here, because they make them “slow” rather than because it is cowardice). In short, the Dothraki have a general contempt for armor and for those who wear it.
The unarmored ‘barbarian’ who attacks heedless of his peril, all fury and offense, no reason or defense (this is a part of the ancient form of the Fremen Mirage), is a literary trope that goes back at least to the Greeks and Romans (::deep breath:: Plut. Marcellus 8; Dionysius 14.9.2; Diodorus 14.9-10; 5.30.3; Liv. 7.10.7-10; App. Gal. 6; Plb. 2.30.1 and 3.114.4 but cf. Liv. 22.46.6; Caes. B.G. 4.1; in artwork note J.R. Marszal, “Ubiquitous Barbarians: Representations of the Gauls at Pergamon and Elsewhere” in From Pergamon to Sperlonga, eds. N.T. de Greummond and B.S. Ridgeway (2000)). In some cases the lack of armor (or clothing) seems to have been accurate and in some it was not (by the by, the Gauls with their supposed barbarian contempt of armor invented mail, probably the single most successful pre-modern armor technology); the idea that these ‘barbarians’ were madly reckless was never accurate – the Gauls used effectively the same full-size body-shield the Romans did which shows a real concern for personal protection! That same language – the irrational, ‘swarming’ natives, heedless of danger or death – reappears in later European military writing, particularly in the early modern and after (which is to say, during European imperial expansion) as a racist marker of non-European inferiority (on this, note P. Porter, Military Orientalism: Eastern War Through Western Eyes (2009), esp. 68ff where he notes the same tropes were applied to Native Americans, which – see below at Stagecoach for the almost certain line of connection where this trope gets to A Song of Ice and Fire).
(And yes, by the by, more than a few of the subjects of this bad old trope seem to have, at one point or another, adopted it in the way the Dothraki do in A Song of Ice and Fire – attributing weakness and cowardice to the ‘westerners’ and their supposed casualty aversion. Both sides of the stereotype are nonsense and states that have made strategic decisions based on those tropes have almost invariable faced disaster. That said, for reasons that will be obvious below, actual nomads do not adopt this position because they haven’t the population to sustain it.)
In short, this is a very old literary trope repurposed as a still-old hoary racist stereotype. It is also hot nonsense, both generally and as applied to our horse-borne nomads.
The popular image of the Great Plains Native America is unarmored, of course, but that image fundamentally formed in the late 19th century, when – after centuries of the development of gunpowder weapons – everyone was unarmored. A longer view shows that Plains Native Americans were perfectly capable of both developing or adopting defensive measures which worked. And to get a full sense of that, we need to outline the major phases of the changing warfare on the Great Plains.
F.R. Secoy (op. cit.) essentially breaks warfare into four phases, which happen at different times in different places, based on if they have horses, guns, both or neither. Because horses entered the Great Plains from the South (via the Spanish) but firearms entered the region from the North (via the British and the French, the Spanish having prohibited gun-sales to Native Americans) and spread out from there, for a brief time many of these systems were active on the Plains at once, as both guns and horses diffused through the region.
In the pre-horse, pre-gun phase (described by McGinnis as well, op. cit., 8-9), battles consisted of long-range missile exchanges between warriors who stood behind large shields which protected their whole bodies. Native American warriors in this system also wore armor, heavy leather coats, laminated in multiple layers using thick hide with glue that was sometimes mixed with sand or gravel (one more example of how ‘leather armor’ is almost always hardened leather armor, not modern clothing-leather). Some of this armor may have been effectively quilted leather as well. Clearly, there was plenty of concern about survivability here.
Both guns and horses were apt to disrupt this system. Horses allowed attackers to rapidly close the distance between the two opposing lines of shield-protected foot-missile-warriors, causing the shield-lines to drop away (though smaller shields, used on horseback to ward off arrows and blows were still used) and for both sides to seek instead the mobility of mounted fighting. That was not the end for armor though, because contact with a supply of horses meant contact with the Spanish, and the Apache at least swiftly adopted some of the Spanish methods of making leather ‘buff coats’ into their own armor tradition and copied the shape and pattern of the buff coat itself (while often still making the material using their own tradition). As Secoy notes (op. cit., 18-20), our sources are quite clear that these forms of armor (both original and Spanish-influenced Native armors) were quite effective at resisting the archery fire that dominated both the pre-horse, pre-gun system and the post-horse, pre-gun system.
Meanwhile on the Northern Plains, while the horse had not yet arrived, firearms had, and these had different effects. Firearms spelled the end of the armor once they became common enough, since there was no protection which could resist bullets; some shield use survived, since arrows remained fairly common as well. But this didn’t lead to suicidal warfare. Instead – as had happened on the East Coast as well, Native Americans adapted their warfare to the increased lethality of firearms (on this, note Lee, op. cit. above) by mostly avoiding pitched battles as they became too lethal (which, by the by, the relatively low lethality of pitched battles pre-gunpowder is often taken to mean that Native North American warfare in general was low-lethality; this is wrong. As with most forms of non-state warfare, most of the killing happened in surprise raids and ambushes, which could be extremely lethal and were still quite common).
Once the horse and the firearm were both in wide use in an area, warfare shifted again. War parties became smaller, faster moving and more reliant on surprise (essentially an extension of the raiding-focus of the pre-horse, post-gun system to the high mobility horses supplied). Infantry battle dropped away entirely because it was too lethal and resulted in casualties that low-population density nomads could not sustain (the contrast with the much higher population-density agrarian United States, which was self-immolating in massively costly massed-infantry engagements during the American Civil War, 1861-1865, at exactly this time is striking). These are fairly big, noticeable changes in warfare patterns!
In short, the tactics used in all four of these systems were conditions by casualty aversion, which makes a lot of sense in the context of a low-population density society which simply cannot afford massive losses.
Meanwhile the Mongols were quite fond of armor, though it is clear that they required access to the products of agrarian economies to get it. That same Khitan regulation I noted above required soldiers to possess nine pieces of iron armor, along with barding (that is, armor) for their horses. John de Plano Carpini describes the use of thins trips of leather and hide, bouind by cord to create a scale of lamellar horse barding. As May notes (op. cit., 53) the Mongols tended to prefer lamellar armors (that is, armors of overlapping rectangular plates attached to each other rather than to a backing) of either hardened leather or iron because these were more effective at stopping arrows than mail. The Mongols also seem to have really liked pointed conical helmets (the Turks did too) and seem to have contributed to their spread. There is actually a fair amount of evidence that the later European brigandine was a Western European adaptation of steppe lamellar armors, mediated through Eastern Europe. In short, Mongol armor (which again, is generally not being produced by them on the steppe but produced for them by the agrarian societies, which in some cases involved violently moving those craftsmen to where the Mongols needed them) was so good that it was quickly adopted in Europe when it arrived.
Indeed, looking broadly at not only the Mongols, but also Turkic steppe nomads (Mamluks, the Seljuks, the Ottomans, etc), incorporating the heavier armor made possible by agrarian societies and their metal production industries seems to have been a fairly high priority for nomads moving into settled zones more generally. And I should note that while the Mongols preferred lamellar to mail (as did other Steppe nomads), they absolutely would and did use mail if mail was what was available; the Turks and the Timurids both made extensive use of mail as well as lamellar, scale and so on. While armor might be rare on the Steppe due to its expense and the relative inability to produce good armor locally, it was valuable and valued and nomads who ‘made good’ worked to acquire it.
In short, if the Dothraki followed the model of either the Great Plains Native Americans or Steppe nomads, far from holding armor in contempt, we’d expect a wealthy leader like Drogo to have made serious efforts to either acquire effective foreign armor both to protect himself and his immediate retainers. Remember that providing armor for your followers would not merely improve their combat effectiveness, but provide an opportunity to reward loyalty and more tightly bind those followers to you.
But of course that wouldn’t be badass and violent enough, so instead Martin deploys the bad old trope of the irrational, mindless barbarian careless to his own peril, going into battle unarmored.
I know we have dealt with this distinction before in a number of other places, but I want to make sure we are all on the same page here, briefly. Tactics concern the how of warfare at the small scale; how a battle is fought. Operations concern how armies are moved and thus where a battle is fought. Strategy concerns the ends for which a war is waged in the first place and thus why battles are fought.
Each of these levels is a category of analysis, but of course not every general starts at first principles when going into a conflict. Instead, wars are often waged according to traditional systems of norms and expectations. That said, when you dig in to those systems of norms and expectations, the basic correlation of strategic ends to the means of operations and tactics generally emerge (if not the least because polities which fail to coordinate these things tend not to be permitted to play the game for very long).
Since strategy dictates operational concerns, which in turn dictate tactical concerns, we’ll move in that order and so begin with the question what does Dothraki warfare aim to achieve and are those the same ends as nomadic warfare on the Great Plains and the Eurasian Steppe?
Assessing the strategy of Dothraki warfare is tricky, because while we spend a good chunk of the story near a Dothraki leader, strategic aims are usually not discussed with our viewpoint characters. Still we see enough of how Dothraki khalasars function to get a sense of the general aims of Dothraki warfare.
First, it seems that the Dothraki Sea is essentially a war of all against all. As Daenerys notes once she is weakened, “the first khalasar they met would swallow up her ragged band, slaying the warriors and slaving the rest” (ACoK, 142). And indeed, we see this when Drogo’s and Ogo’s khalasars raided the same town; Drogo does exactly that, killing the adult men and enslaving the rest of Ogo’s defeated khalasar (AGoT, 556), while Pono’s khalasar scatters others before it as it moves as they rush to get out of the way (ADwD, 113). No khals appear to be allied with each other, there are no tribal confederations of smaller units; merely a collection of khalasars, each led by a khal, all at war with all of the others all of the time except when in Vaes Dothrak. Of course as already mentioned, the agrarian peoples on the edges of the Dothraki Sea are also subjected to this treatment, unless they are either key trade partners or buy the Dothraki off.
The main aim of this functionally total conflict is the capture of slaves. As we’ve discussed previously, the Dothraki don’t herd livestock, and horse-raiding isn’t ever prominent in the text either. Indeed, when attacking the Lhazareen, the Dothraki kill all of the livestock and leave their bodies to rot in the fields (AGoT, 555), instead, they take slaves. Of the Dothraki captives, we are told they are only “the women and children of Ogo’s khalasar;” the captive Lhazareen include “only a few men among them, cripples and cowards and grandfathers” (AGoT, 555-6); so the slaves in question are mostly women, children and the elderly (a strange choice since these are, historically speaking, the least valuable classes of people to enslave; most enslaved workers were forced to do agricultural or industrial labor for which adult males were typically considered the most suitable by far).
Apart but strangely, apart from a few domestics, the Dothraki have apparently no use for these enslaved people internally. Captive women evidently are not able to become wives or bear legitimate sons (AGoT, 559) and without any kind of domestic production (because of the non-existent subsistence model discussed last time) it is hard to imagine what work large numbers of enslaved persons would be forced to do. Instead, we are repeatedly told these masses of slaves are not incorporated into the khalasar, but traded away to the cities of Slaver’s Bay in exchange for ‘gifts’ (ADwD, 662) or left in Vaes Dothrak.
Crucially, in the description of the movements of khalasar, it is clear they do not fight over territory. Drogo’s khalasar cuts all of the way from Pentos over effectively the entire length Dothraki Sea to Vaes Dothrak. He then plans to head to the Jade Sea, which would mean covering the entire height of the Dothraki Sea (since Vaes Dothrak is at its northern extent) and going even further East, but instead cuts south to the territory of the Lhazareen. The movements of the other khalasars (ADwD, 113, 662) are like this as well. The Dothraki essentially cruise the empty Dothraki Sea like actual ships on the actual sea, without any concern for territory. At no point does any group of Dothraki get angry with any other group of Dothraki for territorial trespass.
This bears little resemblance to the strategic concerns of historical nomads. As a direct consequence of failing to understand the subsistence systems that nomads relied on, Martin has also rendered their patterns of warfare functionally unintelligible.
The chief thing that nomads, both Great Plains Native Americans and Eurasian Steppe Nomads used violence to secure control of is the one thing the Dothraki never do: territory. To agrarian elites (who write most of our sources) and modern viewers, the vast expanses of grassland that nomads live on often look ’empty’ and ‘unused,’ (and thus not requiring of protection), but that’s not correct at all. Those ’empty’ grasslands are very much in use; the nomads know this and are abundantly willing to defend those expanses of grass with lethal force to keep out interlopers. Remember: the knife’s edge of subsistence for nomads is very thin indeed, so it takes only a small disruption of the subsistence system to push the community into privation.
For the Eurasian Steppe nomad, the grass that isn’t near their encampment is in the process of regrowth for the season or year when it will be near their encampment and need to support their herds. Allowing some rival nomadic group to move their sheep and their horses over your grassland – eating the essential grass along the way – means that grass won’t be there for your sheep and your horses when you need it; and when the sheep starve, so will you. So if you are stronger than the foreign interloper, you will gather up all of your warriors and confront them directly. If you are weaker, you will gather your warriors and raid the interloper, trying to catch members of their group when they’re alone, to steal horses and sheep (we’ll come back to that); you are trying to inflict a cost for being on your territory so that they will go away and not come back.
The calculus for nomadic hunters like the Great Plains Native Americans is actually fairly similar. Land supports bison, bison support tribal groupings, so tribal groups defend access to land with violent reprisals against groups that stray into their territory or hunt ‘their’ bison. And of course the reserve is true – these groups aren’t merely looking to hold on to their own territory, but to expand their subsistence base by taking new territory. Remember: the large tribe is the safe tribe; becoming the large tribe means having a larger subsistence base. And on either the plains or the steppe, the subsistence base is fundamentally measured in grass and the animals – be they herded sheep or wild bison – that grass supports. Both Secoy and McGinnis (op. cit.) are full of wars of these sorts on the Great Plains, where one group, gaining a momentary advantage, violently pushes others to gain greater territory (and thus food) for itself. For instance, Secoy (op. cit., 6-32) discusses how access to horses allowed the Plains Apache to rapidly violently expand over the southern Plains in the late 17th century, before being swept off of them by the fully nomadic Ute and Comanche in the first third of the 18th. As McGinnis notes (op. cit., 16ff), on the Northern Plains, prior to 1800 it initially was the Shoshone who were dominant and expanding, but around 1800 began to be pushed out by the Blackfoot, who in turn would, decades later, be pushed by the expanding Sioux.
This kind of warfare is different from the way that settled, agrarian armies take territory. Generally, the armies of agrarian states seek to seize (farm-) land with its population of farmers mostly intact and exert control both over the land and the people subsequently in order to extract the agricultural surplus. But generally (obviously there are notable exceptions) nomads both lack the administrative structures to exert that kind of control and are also very able to effectively resist that sort of control themselves (it is hard for even nomads to tax nomads), making ’empire building’ along agrarian lines difficult or undesirable (unless you are the Mongols). So instead, polities are trying to inflict losses (typically more through raiding and ambush than battle). Since rivals will tend to avoid areas that become unsafe due to frequent raiding, the successful tribe can essentially push back an opposing tribe with frequent raids. In extreme circumstances, a group may feel threatened enough to get up and move entirely – which of course creates conflict wherever they go, since their plan is to disposes the next group along the way of their territory.
Within that security context, larger scale groupings – alliances, confederations, and super-tribal ‘nations’ – are common. On the Eurasian Steppe, such alliances tended to be personal, although there was a broad expectation that a given ethnic grouping would work together against other ethnic groupings (an expectation that Chinggis actually worked very hard, once he became the Great Khan of a multi-ethnic ‘Mongol’ army, to break up through the decimal organization system; this reorganization is part of what made the Mongol Empire so much more successful than previous Steppe confederations). Likewise, even a cursory look at the Native Americans of the Great Plains produces both a set of standard enmities (the Sioux and the Crow, for instance) but also webs of peace agreements, treaties, alliances, confederations and so on. The presence of British, French, Spanish and American forces (both traders and military forces) fit naturally into that system; the Plains Apache allied with the Spanish against the Comanche, the Crow with the United States against the Sioux and so on. Such allies might not only help out in a conflict, but also deter war and raiding because their strength and friendship made lethal retaliation likely (don’t attack someone allied to Chinggis Khan and expect to survive the experience…).
Exactly none of that complexity appears with the Dothraki, who have no alliances, no peace agreements, no confederations and no territory to attack or defend. Instead, the Dothraki simply sail around the grass sea, fighting whenever they should chance to meet. Which brings us to:
The other strategic aim nomads might fight over is for the acquisition of some kind of movable good, which is to say raiding for stuff. Because all of the warriors (which is generally to say all of the free adult males) of these societies are mounted and because they have a subsistence system which allows rapid, relatively along distance movements (often concealed; remember that Mongols need not light any camp fires), nomads make fearsome raiders, able to strike, grab the things they are looking for and quickly retreat before a counterattack can be mobilized. That goes just as well for raiding each other as it does for raiding the farmers at the edges of the grasslands.
But what are the things here that they are aiming to get? It depends on the targets; nomadic raids into the settled zone generally aim to capture the goods that agrarian societies produce which nomadic societies do not: stocks of cereal crops, metal goods and luxury goods. But most nomadic raiding was directed against other nomads, seeking to acquire either people or animals.
On the Great Plains, the animals in question were invariably horses; the act of stealing, or ‘cutting out’ a horse gives McGinnis part of the title of his book (Counting Coup and Cutting Horses) and raids for horses dominate both McGinnis and Secoy’s discussion of Plains Native American warfare. Horses were, after all, a scarce commodity which only percolated into the Great Plains from the South (and which could only be raised in quantity in its southern reaches), but which all tribes required both to hunt and fight effectively. Stealing enemy horses thus both strengthened your tribe while weakening your enemies, both in military and subsistence terms. The Mongols also engaged in quite a lot of raiding for horses, but also – in a pastoral subsistence system – a lot of simple cattle rustling as well (e.g. Ratchnevsky, op. cit., 28-31).
Raiding for people is more complex, but undeniably part of this system of warfare. But crucially this raiding was generally not for slave-trading (though there are exceptions which I discussed last time), but instead incorporative raiding. What I mean by that is that the intent in gaining captives in the raid was to incorporate those captives, either as full or subordinate members, into the nomadic community doing the raiding. Remember: the big tribe is the safe tribe, so incorporating new members is a good way to improve security in the long run.
On the Eurasian Steppe, incorporated captives became the ötögus bo’ol ‘bonded serfs’ that we mentioned previously (Ratchnevsky, op. cit., 12-4). Unlike warfare on the Great Plains, it seems possible for the bo’ol to include adult men, either captured or sold (by destitute parents) as children or else taken as prisoners when their tribe or clan was essentially dissolved by being conquered in war. Indeed, in his own conquests, Chinggis only decreed the annihilation of one tribe, the Mongol’s traditional enemies, the Tatars – there he ordered the death of any Tatar male taller than the linchpin of an oxcart (May, Mongols, 12). In other cases, it is clear that the incorporation of defeated nomad warriors into the successful tribe was fairly normal, though raids to capture women and children (also for incorporation) were just as common. Bride abduction in particular was very common on the Steppe, as Ratchnevsky notes (op. cit., 34-5).
The incorporation of males was far less common in Great Plains Native American warfare, but the capture of women and children to enhance tribal strength in the long term was a core objective in raiding. McGinnis (op. cit., 42-3) notes how the Crow, after suffering a massive defeat in the early 1820s which resulted in the deaths of many warriors and the capture of perhaps several hundred women and children, steadily built their tribe back up over the following decades with an intentional strategy of capturing women and children from their enemies. As McGinnis (op. cit., 24) notes, women captured in this way might be married into the capturing tribe, adopted into it, or sometimes kept as an enslaved laborer (under quite bad conditions). Adult males, by contrast, were almost always killed; unlike on the Steppe, the incorporation of formerly hostile warriors doesn’t seem to have been considered possible (though one wonders if this would have become cultural practice given enough time; both McGinnis and Secoy note how the increasing lethality of warfare post-gun/horse led to slow population decline overall, which may, had the system run without outside interference long enough, led to the emergence of norms more closely resembling the Eurasian Steppe. We should keep in mind that the Eurasian horse-system had many centuries to sort itself out, whereas the North American horse-system was essentially strangled in its crib).
Of course, taken together with the previous discussion of territorial warfare, we can see that all of these raids have a double purpose: they both aim to acquire resources (horses, sheep, humans) and at the same time inflict damage on an opponent with the long-term goal of forcing that enemy to move further away, opening their pastures or hunting grounds for exploitation by the victorious tribe. Thus in the long-term, each successful raid is intended to build a sense of threat which eventually results in territorial gains (though in cases of real power asymmetry, the long term could come very rapidly; people aren’t stupid and if you are being raided by a clearly superior opponent, you are likely to move on before you lose everything of value).
Squaring the ugly reality of nomadic raiding with Martin’s depiction is tricky. On the one hand, a raid in which exceptional victory results in enemy women and children taken captive and fit adult males slain fits within either the Great Plains Native American or Steppe nomad military tradition. On the other hand, the immediate declaration by Drogo’s men that female captives taken this way are not marriageable (AGoT, 559; the idea is treated as laughable) and the killing of all of the very valuable livestock (which, even if the Dothraki are not herdsmen, these animals could be eaten, or quite easily driven to a place where they could be sold or traded for other resources, like metalwork) suggests that Martin has not understood why those raids happened. Instead, it seems like his imagination is only able to view these raids from the perspective of the settled people on the receiving end.
Instead, Martin’s understanding of Native American warfare seems not conditioned by any actual Native Americans, but rather by Hollywood depictions of Native Americans during the Hollywood ‘Golden Age’ which were in turn conditioned by sensational accounts of Western settlers who themselves didn’t understand how Native American warfare worked on the Great Plains. As we will see, the Game of Thrones showrunners took that unfortunate subtext when making the show itself, and turned it into actual text.
The Preposterous Tactics of the Dothraki
We do not see the Dothraki engage in large-scale warfare in the books; we see the aftermath of such fighting (AGoT, 555ff) or it occurs ‘off-screen’ (ASoS, 487), but we do not see it. The closest we get is Jorah’s description of them, that they are “utterly fearless…[they] fire from horseback, charging or retreating, it makes no matter, they are full as deadly…and there are so many of them” AGoT, 325-6). Evidently they also scream on the attack, since their warriors are repeatedly called ‘screamers.’
As a description, it is hard for this to be very much wrong because it is so very vague, but the attentive reader will note that Jorah’s assertion that there are ‘so many’ must be wrong for either Eurasian Steppe Nomads or Great Plains Native Americans, both of whom were routinely outnumbered by settled enemies, often dramatically so. Let’s put a pin in that, though, because of course while Martin gives only vague description of Dothraki warfare, the show, Game of Thrones, shows it to us on screen quite vividly.
We see a bit of Dothraki warfare in S6E9 when Daenerys’ Dothraki charge down the Sons of the Harpy at Mereen, but the really sustained look at how they fight has to wait for S7E4 and the Loot Train Battle and S8E3 and the Battle of Winterfell, both of which, happily, we have already discussed! In all three cases, the Dothraki do exactly the same thing. They charge, in a pell-mell rush, while giving high-pitched war-calls. While some of the Dothraki may fire arrows on the approach (they have them stand up to do this, which is not how actual Mongols or Native Americans fired from horseback; it looks cool and is stupid, like most of Game of Thrones season 7 and 8), they otherwise charge directly into contact and begin fighting from horseback with their arakhs as the primary weapon.
This is not how horse-borne nomads fought.
As we’ve discussed repeatedly before, the key weapon for Steppe nomads was the bow, shot from horseback at high speed (on this, note May, “The Training of an Inner Asian Nomad Army” JMH 70 (2006) and Mongol Art of War (2007)). Thus the crucial maneuver was the caracole, where the rider approaches the target at high speed, firing arrows as he goes, before making an abrupt turn (it is actually the turn that is technically called a caracole, but the whole tactic goes by this name) and retreating, before trying again. Pulling this tactic off en masse required a great deal of both individual skill at horsemanship and archery, but also quite a lot of group cohesion and coordination, since a collision of horses at speed is very likely to be fatal for everyone – humans and horses – involved.
This tactic can then be repeated – charge and retreat, charge and retreat – until the psychological toll on the defender becomes too great and they either break and retreat or else charge out to try to catch ‘retreating’ nomads. In either case, it was at that moment when the Steppe nomads could press home and destroy the disorganized enemy. These tactics were brutally effective, but they were also a necessary casualty control measure. Shock combat – that is massed melee combat in close quarters – is simply far too lethal for low-population nomadic societies to sustain in the long-term on the regular (a hoplite battle might result normally in c. 10% casualties for instance (but note this discussion of that figure) – think of what that would mean in a society where 100% of adult males participate in each battle – you’d run out of men pretty quickly!).
And fascinatingly, we can actually see that calculus play out in North America, where the arrival of firearms, which suddenly make pitched ‘missile exchange’ battles (especially on foot) as lethal as shock combat (it seems notable that the introduction of musketry into Old World warfare did not come with a significant increase or decrease in battlefield lethality, at least until the rifled musket – on that, see B. Gibbs, The Destroying Angel (2019), but also noteE.J. Hess, The Rifle Musket in Civil War Combat: Reality and Myth (2008)), the pitched battle vanishes. It was simply too lethal to be a viable option in the long term for societies with low population density and very high military participation rates.
Instead, the raid came to dominate warfare on the Great Plains, with mass-casualty events generally being restricted to situations where a raiding party caught an enemy group unawares (McGinnis, op. cit., 45-6, 57-9). To be clear, that’s not to say the Great Plains Native Americans were peaceful, after all the goal of all of this raiding was to cause one of those rare mass-casualty surprise attacks and – as McGinnis notes again and again, warfare was part of the Plains Native American way of life, as the social status of males was directly and powerfully tied to success in war.
In short, the need to keep lethality relatively low is one of the most important factors which shaped nomadic horse-borne warfare, both on the Steppe and on the Great Plains. And here is where I think that even Martin’s description – which could, if read with friendly eyes, be taken as a description of the Steppe caracole described above – falls short: the Dothraki are dangerous because they are so many. But actual nomadic warfare was fundamentally conditioned by the shortage of men created by the low population density of the Steppe or the Great Plains. This weakness could be somewhat made up for by making every male into a warrior, but only if casualty rates remained low. A war of attrition with settled peoples would wear the nomads out quickly, which is why such attritional warfare was avoided (unless you are the Mongols, who use the sedentary armies of conquered states, notably using the armies of Northern China to conquer Southern China; that said, Drogo is clearly not Chinggis Khan or any such sort of Khal-of-Khals).
So where does this model of warfare come from? Well, when it comes to the show, we needn’t actually look far, because the creators tell us. The director of the episode, Matt Shakman, noted in an interview that his primary reference for the Dothraki charge was John Ford’s Apache attack in his 1939 film Stagecoach (you can see the scene he means here). And in the S7 special feature, “Anatomy of a Scene: The Loot Train Attack,” David Benioff notes that the charge “definitely got a bit of that western feel” while VFX producer Steve Kullback says, of the battle, it’s “sort of like Cowboys and Indians.”
In Stagecoach (1939), the Apache aren’t a real humanized culture, but an elemental force of destruction. Their charge at the titular stagecoach is essentially mad and heedless of all losses (in the same featurette, Camilla Naprous, Game of Throne‘s horse master, describes the Dothraki as “they’re just these absolute mad men on horses,” in case you thought that connection was only subtext). The position of ‘Indians’ as particularly ‘rapey’ is also explicit in Stagecoach, where the one of the white male defenders of the coach saves his last bullet to spare the one woman, Mrs. Mallory, from being captured and raped by the approaching cavalry (the concern about white women being raped by non-white men being a paramount fixation of early American film; see also The Birth of a Nation (1915); or, you know, don’t.) And the tactics (or lack thereof) of the Dothraki, charging madly forward with no order or concern for safety, also map neatly on to Stagecoach‘s Apache attack (and not on to actual Apache attacks).
I don’t think this lazy use of old Western tropes is limited to merely the show, however. Having written this far, I find myself convinced that there is a longer article or perhaps a video-essay waiting to be written by a different sort of scholar than myself – that is, a film historian – on how Martin’s depiction of the Dothraki and their world is fundamentally rooted in the racist tropes of the Hollywood Western and its portrayal of Native Americans in a frontier environment where, as Sergio Leone put it, “life has no value.” Quite a lot of parallels with Martin’s Dothraki emerge after even a brief overview of the representation of Native Americans in film. The emphasis on taking captives (especially white women) to no apparent purpose besides sexual violence, the distinctive ‘screaming’ of Dothraki warfare (which, yes, Native Americans used a range of intimidating war cries, but so did basically everyone else in the pre-modern world, so why are the Dothraki the only ones who do it in Westeros?), it’s lack of tactics or order, and – as we’ve discussed already – the grossly simplified form of dress all seem to have their roots in racist Hollywood depictions of Native Americans. The Dothraki Sea is, essentially a ‘Cavalry and Indian Story‘ with the cavalry removed.
That is not a pure creation of Benioff and Weiss. The show simply takes that subtext and makes it text.
That makes it a good time, here at the end, to take stock. As I’ve noted in each of these posts, the fundamental claim we are evaluating here is this one, made baldly by George R.R. Martin:
We may, I think, now safely dismiss this statement as false. What we have found is that the Dothraki do not meaningfully mirror either Steppe or Plains cultures. They do not mirror them in dress, nor in systems of subsistence, nor in diet, nor in housing, nor in music, nor in art, nor in social structures, nor in leadership structures, nor in family structures, nor in demographics, nor in economics, nor in trade practices, nor in laws, nor in marriage customs, nor in attitudes towards violence, nor in weapons, nor in armor, nor in strategic way of war, nor in battle tactics.
We might say he has added ‘dashes’ of pure fantasy until the ‘dash’ is the entire soup, but the truth is clearly the reverse: Martin has sprinkled a little bit of water on a barrel of salt and called it just a dash of salt. There is no historical root source here, but instead pure fantasy which – because racist stereotypes sometimes connect, in thin and useless ways, to actual history – occasionally, in broken-clock fashion, manages to resemble the real thing.
It seems as though the best we might say of what Martin has right is that these are people who are nomads that ride horses and occasionally shoot bows. The rest – which as you can see from the list above there, is the overwhelming majority – has functionally no connection to the actual historical people. And stunningly, somehow, the show – despite its absolutely massive budget, despite the legions of scrutiny and oversight such a massive venture brings – somehow is even worse, while being just as explicit in tying its bald collection of 1930s racist stereotypes to real people who really exist today.
Instead, the primary inspiration for George R.R. Martin’s Dothraki seems to come from deeply flawed Hollywood depictions of noamdic peoples, rather than any real knowledge about the peoples themselves. The Dothraki are not an amalgam of the Sioux or the Mongols, but rather an amalgam of Stagecoach (1939) and The Conqueror (1956). When it comes to the major attributes of the Dothraki – their singular focus on violent, especially sexual violence, their lack of art or expression, their position as a culture we primarily see ‘from the outside’ as almost uniformly brutal (and in need of literally the whitest of all women to tame and reform it) – what we see is not reflected in the historical people at all but is absolutely of a piece with this Hollywood legacy.
But Martin has done more damage than simply watching The Mongols (1961) would today. He has taken those old, inaccurate, racially tinged stereotypes and repackaged them, with an extra dash of contemporary cynicism to lend them the feeling of ‘reality’ and then used his reputation as a writer of more historically grounded fantasy (a reputation, I think we may say at this point, which ought to be discarded; Martin is an engaging writer but a poor historian) to give those old stereotypes the air of ‘real history’ and how things ‘really were.’ And so, just as Westeros became the vision of the Middle Ages that inhabits the mind of so many people (including quite a few of my students), the Dothraki become the mental model for the Generic Nomad: brutal, sexually violent, uncreative, unartistic, uncivilized.
And as I noted at the beginning of this series, Martin’s fans have understood that framing perfectly well. The argument given by both the creators themselves, often parroted by fans and even repeated by journalists is that A Song of Ice and Fire‘s historical basis is both a strike in favor of the book because they present a ‘more real’ vision of the past but also a flawless defense against any qualms anyone might have over the way that the fiction presents violence (especially its voyeuristic take on sexual violence) or its cultures. No doubt part of you are tired of seeing that same ‘amalgam’ quote over and over again at the beginning of every single one of these essays, but I did that for a reason, because it was essential to note that this assertion is not merely part of the subtext of how Martin presents his work (although it is that too), but part of the actual text of his promotion of his work.
And it is a lie. And I want to be clear here, it is not a misunderstanding. It is not a regrettable implication. It is not an unfortunate spot blind-spot of ignorance. It is a lie, made repeatedly, now by many people in both the promotion of the books and the show who ought to have known better. And it is a lie that has been believed by millions of fans.
One thing that I hope is clear from this treatment is just how trivial the amount of research I’ve done here was. Certainly, it helped that I was familiar with Steppe nomads already and that I knew who to ask to be pointed in the direction of information. Nevertheless, everything I’ve cited here is available in English and it is all relatively affordable (I actually own all of the books cited here; thanks to my Patrons for making that possible, especially since getting materials from the library is slower in the days of COVID-19; nevertheless, the point here is that they are not obscure tomes). Much of it – Ratchnevsky on Chinggis Khan, Secoy and McGinnis on Great Plains warfare – were already available well before the 1996 publication of A Game of Thrones. 1996 was not some wasteland of ignorance that might have made it impossible for Martin to get good information! For an easy sense of what a dedicated amateur with film connections might have learned in 1996, you could simply watch Ken Burn’s The West, which came out the same year. I am not asking Martin to become a historian (though I am asking him to stop representing himself as something like one), I am asking him to read a historian.
Instead of doing that basic amount of research, or simply saying that the peoples of Essos were made up cultures unconnected with the real thing, Martin and the vast promotional apparatus at HBO opted to lie about some real cultures and then to put hundreds of millions of dollars into promoting that lie.
And I want to be clear, these are real people! I know, depending on where you live, ‘Mongols’ and ‘Sioux’ and ‘Cheyenne’ may feel as distant and fanciful as ‘Rohirrim’ or ‘Hobbits’ or else they may feel like ‘long-lost’ peoples. But these were real people, whose real descendants are alive today. And almost all of them face discrimination and abuse, sometimes informally, sometimes through state action, often as a result of these very lingering racist stereotypes.
In that context, declaring that the Dothraki really do reflect the real world (I cannot stress that enough) cultures of the Plains Native Americans or Eurasian Steppe Nomads is not merely a lie, but it is an irresponsible lie that can do real harm to real people in the real world. And that irresponsible lie has been accepted by Martin’s fans; he has done a grave disservice to his own fans by lying to them in this way. And of course the worst of it is that the lie – backed by the vast apparatus that is HBO prestige television – will have more reach and more enduring influence than this or any number of historical ‘debunking’ essays. It will befuddle the valiant efforts of teachers in their classrooms (and yes, I frequently encounter students hindered by bad pop-pseudo-history they believe to be true; it is often devilishly hard to get students to leave those preconceptions behind), it will plague efforts to educate the public about these cultures of their histories. And it will probably, in the long run, hurt the real descendants of nomads.
But this is exactly why I think it is important for historians to engage with the culture and to engage with depictions like this. Because these lies have consequences and someone ought to at least try to tell the truth. With luck, even with my only rudimentary knowledge, I have done some of that here, by presenting a bit more of the richness and variety of historical (and in some cases, present-day) horse-borne nomadic life, in both North America and Eurasia.
Because there is and was a lot more to nomads than just ‘that Dothraki horde.’