Fireside Friday, April 22, 2022

Fireside this week! We’re in the last few weeks of the semester, but semesters tend to ‘crescendo’ rather than ‘wind down’ so there has been a lot going on. I’ll probably be posting a gap week for next week (Friday, April 29) because I’ll be at the annual meeting of the Society for Military History in Fort Worth, TX. I’m not presenting this year, but I am chairing a panel on ‘Issues of Governance, War and Identity from Rome to Vietnam.’

For this week’s musing, I want to take on a bit of a lighter topic: the size of science fiction space armies (and to a lesser extent, fleets). It has long been observed that Sci-Fi writers have no sense of scale and that is certainly a fair observation. Indeed, when it comes to the size of armies, most writers seem to have little sense of scale, but in galaxy-spanning space operas one can be several orders of magnitude off of the correct scale. Nevertheless, I want to caveat some assumptions that various science fiction civilizations ought to have trillions of combat troops or many thousands of warships.

The first issue is what in military parlance is called the ‘tooth to tail’ ratio. This is the ratio of the number of actual combat troops (the ‘tooth’) to logistics and support personnel (the ‘tail’) in a fighting force. Note that these are individuals in the fighting force – the question of the supporting civilian economy is separate. The thing is, the tooth to tail ratio has tended to shift towards a longer tail over time, particular as warfare has become increasingly industrialized and technical.

The Roman legion, for instance, was essentially all tooth. While there was a designation for support troops, the immunes, so named because they were immune from having to do certain duties in camp, these fellows were still in the battle line when the legion fought. The immunes included engineers, catapult-operators, musicians, craftsmen, and other specialists. Of course legions were also followed around by civilian non-combatants – camp-followers, sutlers, etc. – but in the actual ranks, the ‘tail’ was minimal.

You can see much the same in the organization of medieval ‘lances,’ – units formed around a single knight. The Burgundian ‘lance’ of the late 1400s was composed of nine men, eight of which were combatants (the knight, a second horsemen, the coustillier, and then six support soldiers, three mounted and three on foot) and one, the page, was fully a non-combatant. A tooth-to-tail ratio of 8:1. That sort of ‘tooth-heavy’ setup is common in pre-industrial armies.

The industrial revolution changes a lot, as warfare begins to revolve as much around mobilizing firepower, typically in the form of mass artillery firepower as in mobilizing men. We rarely in our fiction focus on artillery, but modern warfare – that is warfare since around 1900 – is dominated by artillery and other forms of fires. Artillery, not tanks or machine guns, after all was the leading cause of combat death in both World Wars. Suddenly, instead of having each soldier carry perhaps 30-40kg of equipment and eat perhaps 1.5kg of food per day, the logistics concern is moving a 9-ton heavy field gun that might throw something like 14,000kg of shell per day during a barrage1, for multiple days on end. Suddenly, you need a lot more personnel moving shells than you need firing artillery.

As armies motorized after WWI and especially after WWII, this got even worse, as a unit of motorized or mechanized infantry needed a small army of mechanics and logistics personnel handling spare parts in order to stay motorized. Consequently, tooth-to-tail ratios plummeted, inverted and then kept going. In the US Army in WWI, the ratio was 1:2.6 (note that we’ve flipped the pre-industrial ratio, that’s 2.6 non-combat troops for every front line combat solider), by WWII it was 1:4.3 and by 2005 it was 1:8.1. Now I should note there’s also a lot of variance here too, particularly during the Cold War, but the general trend has been for this figure to continue increasing as more complex, expensive and high-tech weaponry is added to warfare, because all of that new kit demands technicians and mechanics to maintain and supply it.

Why do I think this is relevant to science fiction armies?

Because there’s no real reason to suppose that trend wouldn’t continue in most science fiction universes. First of, we’ve already gone over why I thin even in a spare-faring context, ground operations might not be dead letter at all. That said, an army structured for orbit-to-ground combat operations is going to have wild logistics demands, even if we assume lots of exotic technology solving things like, for instance, the prohibitive cost of getting military equipment to and from orbit. You’d be looking, at least, an army with armored and mechanized forces (since it needs to have a firepower advantage once it is on the ground against defenders) that needs to be air-droppable. In most cases, armies in science fiction appear to be almost entirely air mobile (e.g. the marines of Halo with their Pelican dropships or the Clone Army in the Star Wars prequels; both appear to be 100% air mobile, including armor), which in turn means fueling, repairing and operating a massive fleet of air mobility platforms both for your troops and also for heavy weaponry. Moving just the infantry component of a single infantry brigade combat team would, for instance, demand something like 300 of Halo’s Pelican dropships (or around 140 V-22 Ospreys if you want a modern platform with similar V/STOL capabilities).

That sort of configuration poses formidable logistics challenges, but it makes a degree of sense. The attacker in such a scenario has a huge mobility advantage already coming from orbit (ships in low orbit circle a planet very fast, enabling them to deorbit basically anywhere on very short notice), but would surrender that advantage if troops once landed couldn’t re-orbit or at least become air mobile quickly. On the flip side the defender is likely to have at least numerical parity if simply because they don’t need to fly through space to get to the battlefield. They also don’t need combat platforms which can be orbit-dropped, so they can employ literally heavier combat systems. Emphasizing mobility, airpower and orbital fires all seem like ways the attacker can level the playing field. But of course then you need to maintain all of that equipment, and stock all of that firepower, though you at least have the advantage that much the ‘tail’ can remain in orbit in space (avoiding the cost of both de-orbiting but also re-orbiting when the operation is done).

At the same time, another trend we’ve seen as the technological and industrial demands of war have increased is that the militarization rate – the percentage of individuals under arms when a country is at war – essentially peaked in WWII and has since declined. Part of this is the cost of the modern military kit: the most effective kinds of fighting are really expensive on a per-soldier basis and so the armies that use them have gotten smaller because a small number of modern-system units is more effective than a large number of static-system units, while at the same time you can only afford a smaller number of modern-system units because they’re so expensive. That has gone hand in hand with increasing dispersion in response to firepower: you need to spread out your systems because enemy firepower is so intense. That limits how many soldiers can operate on a given length of ‘front.’

Contrasting military operations in the Ukraine in WWII to military operations in Ukraine today actually displays some of these effects very clearly, especially when you keep in mind that the USSR of 1940 had a similar, but lower total population to the Russian Federation today (c. 110m to 140m). While neither side has truly ‘filled the frontage’ across the entire border, there is effectively continuous contact along a large front from Kharkiv through the Donbas, despite the forces involved being substantially smaller than the forces involved in the same area in WWII. The dispersion is much higher and thus the density of forces in combat is lower in order to try to avoid the tremendous firepower of the enemy. For cinematic reasons, science fiction shows and movies often show troops at extremely high densities (often closer to late 18th, early 19th century warfare), but in practice dispersion would limit the total number of troops that could be deployed over a given frontage.

In that context, while perhaps a Space Empire could mass-mobilize billions of troops, it’s not clear that they would do so, even in a war where planetary invasions were important combat operations. Instead, they might well assess that the better ‘bang for their buck’ would be a two-tier army: planetary defense forces as a jumbo-sized version of reserves/national guard/territorial defense units, supplemented by a much smaller, much more materiel intensive maneuver army designed for offensive operations like planetary invasions, with the latter soaking up much of the budget as its fancy high-tech systems were expensive to build and maintain.

Now that kind of maneuver army would still almost certainly be larger than what we see in most science fiction settings; probably low millions for a force that plans to be able to seize one planet at a time; much higher for a military that wants to be able to engage in multiple such operations at once. But it wouldn’t necessarily mean billions or trillions of soldiers, either in the tooth or the tail.

Meanwhile, of course, there is a second limiting factor here, which is the degree to which the state can actually access and organize productive resources. The armies of Dune are, for instance, fairly small. The invasion of Arrakis seems to have been done with just 11 ‘legions’ of 30,000 men each (330,000 total; one of these legion are of Sardaukar, the rest Harkonnen), which is not a lot given that it represents essentially a maximum effort by one of the strongest Great Houses in the Imperium. But then the social system of Dune, the faufreluches, doesn’t lend itself to total economic mobilization either. Social class is fairly rigid and inflexible. Meanwhile, a system based on personal hereditary rule and personal relationships (the kind where Dune‘s – the book, to be clear – banquet scene is a useful political exercise; think about what it means about social organization that the banquet is a better use of everyone’s time than mass politics) is going to struggle to mobilize meaningful resources on a planetary scale.

Likewise, the governments in Star Wars (both Republics and the Empire) clearly haven’t been able to extend anything like a modern administrative state over most of the galaxy. Lando is, after all, running a self-governing independent mining city and only gets noticed because his rebel friends get tailed there. Whole sections of the capital planet seem to exist effectively outside of the supervision of the central government and we regularly see planets that clearly exist outside of the meaningful reach of the central government. In short, the vastness of population and physical distance prevents those governments from creating the same kind of total economic and popular mobilizations that smaller states on Earth manage, which is going to limit their military capability. That’s not a crazy notion either; as empires grow larger, their ‘militarization rates’ tend to fall because new conquests cannot be administered as intensively as old ones, because they are culturally different and further away.

A good way of thinking about this is how the different military structures of the ancient Mediterranean (oriented around cities and towns with local civic governments) and much of the medieval Mediterranean (oriented around landed mounted military aristocrats and their retinues) meant that the ability of these societies to mobilize resources for war declined, with most medieval armies in the broader Mediterranean and Europe being substantially smaller than what those areas could produce in earlier centuries. It isn’t entirely silly for a science fiction author to posit that the difficulties in ruling vast expanses of states might enforce similar limits.

And of course all of this goes double for fleets, which are nothing but big, expensive materiel issues, especially when one keeps in mind how absurdly huge many space warships are. Tthe interior space of something like the Battlestar Galactica, for instance, is probably a few hundred times larger than the largest modern warships when calculating for volume or mass. A Star Destroyer is even somewhat bigger than that; one estimate suggests that an Imperial Star Destroyer might mass something like 50 million tons, 500 times the displacement of the world’s largest warships. Given that, in the broader lore, Star Destroyers have to be supported by other smaller, but still very large warships, I am actually not surprised that the Galactic Empire has only a limited number of these platforms and can only concentrate a much smaller number in a single place.

All of which is to say that while I’d expect Future Space Armies to be very large, I don’t think you can cleanly extrapolate from the size of armies in WWII to get to the frontline combat strength of Future Space Armies: the tail is going to be much, much longer and the demands of logistics and resource mobilization may also sharply limit the degree to which a population can be militarized. A lot depends on the sort of society being posited here.

On to recommendations!

Over at Peopling the Past, for April they’ve been focusing on the study of human migration. Megan Daniels offers a historiography (the ‘history of the history’) of how archaeology has been used to study human movements and the shaping assumptions that have often been imported into that study. Catherine Cameron discusses her research on “Unwilling Migrants,” human migration through capture and captivity in the Americas; though I will note that while Cameron describes it as ‘astonishing” to find that most captives were women and children this is actually a very typical pattern in raiding both because adult males are the most able to escape a raid but even moreso because the adult males are killed rather than taken captive.

Meanwhile, Paul Johstono and Michael Taylor recently published a reassessment of the Battle of Pydna (168), the battle that effectively ended the Macedonian kingdom and secured Roman hegemony over Greece, in the journal of Greek, Roman and Byzantine Studies, which is gloriously open-access. Pydna is a tough battle to get to grips with as our sources for it are less than ideal: Livy would normally be a reliable source (deriving most of his details from the text of Polybius, lost for these years) but the surviving corpus of Livy cuts out for much of the early battle; Plutarch offers a complete narrative but much shorter and also less reliable. The source difficulties make a detailed effort in reconstruction like this valuable and well worth a read. If you do look at this, be sure to note the maps at the end, which do a lot to make sense of how the reconstruction is proceeding.

Events in Ukraine continue to show the primary of logistics in war and I’d be remiss, as the ‘orc logistics guy‘ if I didn’t point some of this out. Emily Ferris in Foreign Policy notes that “Russia’s Military Has a Railroad Problem,” a discussion of Russia’s logistics issues stretching from its reliance on rail logistics, which of course will be a familiar problem for those that read last December’s reading recommendation, Alex Vershinin’s “Feeding the Bear: A Closer Look at Russian Army Logistics and the Fait Accompli,” which seems to have largely correctly called the degree to which Russian ground forces would struggle to maintain their logistics beyond the railheads.

This week’s book recommendation is T. Sheppard, Commanding Petty Despots: The American Navy in the New Republic (2022), a history of the early formation of the senior officer corps for the US Navy from 1775 to 1824. The focus of the book is on how the often free-wheeling and startlingly glory-seeking command ethos of the early navy was reformed, albeit in degrees, into an organizational culture which valued professionalism and civilian supremacy over the navy.

Sheppard doesn’t treat this question in abstraction, but rather as a series of concrete narratives following the actions and careers of a progression of navy secretaries and captains. That both serves to make the book very readable – there are no problems with jargon here and the often colorful antics of some of these captains makes for entertaining reading – but it is also a good study into how a succession of leaders can shape organizational culture. Because of the slow communications of the period, the interaction between the ‘petty despots’ (the captains) of the title and the Secretaries of the Navy takes a sort of ‘call and response’ cadence: captains go out and do something and then the Secretary of the Navy has to respond later when they find out; they can’t micromanage their petty despots. Nevertheless, the long series of these calls-and-responses, especially after 1815, slowly established expectations for what command decisions and attitudes would be acceptable and what wouldn’t be. The shift in organizational culture is thus not examined in the abstract, but as a series of concrete decisions: who to praise, who to critique, who to promote, who to cashier, which collectively produce those abstract outcomes.

I find that particularly valuable: there is a tendency to treat questions of command ethos and organizational culture in broad abstract terms, especially when writing for a general reader. That can make sense in a lot of contexts, but it can also make it difficult to understand in a concrete sense what is actually being done to shape or change that organizational culture, to the point that organizational culture can seem impossible to shape or reform. What Sheppard demonstrates is how those ongoing ‘negotiations’ between Secretaries of the Navy, the eventual Navy Board and naval officers themselves steadily shaped and formed a more restrained, deferential and professional naval officer ethos.

  1. At the Somme in 1916, the British artillery fired around 1,050 shells per artillery piece in just about a week; around 150 a day.

280 thoughts on “Fireside Friday, April 22, 2022

  1. This is the only intelligent treatment of this subject I’ve ever seen

    The rest just descend into biggaton wank

      1. Sadly, that’s not how ratios work.

        Looking at some images of gaunt models, I’d estimate that their tails are about 1.5 meters long, but their teeth 20 centimeters at the longest (though with a lot of variation—a surprising number of models don’t have visible teeth). If those figures are accurate, the tooth-to-tail ratio would be 0.2:1.5 = 1:7.5, roughly equivalent to turn-of-the-millennium army.

        1. Sadly, that’s not how ratios work either. From the models, I’d estimate most Tyranids have of the order of 20-50 teeth, and either 0 or 1 tail. That gives a very pre-modern ratio of somewhere between 20:1 and infinity.

    1. Canonically, a single Space Marine chapter is 1000 fighting men. They usually occupy a planet (and secure it for the Imperium.) That implies an absolutely massive tooth-to-tail ratio, with a planet of a billion souls supporting an army of 1000 men. (Of course, not all of those persons on the planet are directly supporting the chapter.)

      Imperial Guard regiments have smaller implied tooth-to-tail ratios, though that’s a little bit more speculative. And a Guard regiment is a fire-and-forget weapon. They’re raised and sent out on campaign and then never reinforced, just consumed. Because The Imperium has reserves.

      1. Except that they are. Gaunt’s Ghosts had the problem that the planet was destroyed, and then the hive for the first batch of reinforcements. Now they merged with another with a functional planet, they are getting them.

        Cain observed of the Valhallan regiment that being merged, it got reinforcements for both while he was there, which really helped, so sometimes they were sparse. But they happened.

        1. Like basically everything about the Imperium, it depends. What does seem to be reasonably reliably true is that regiments, with the notable exception of Cadians, do not tend to return to their homeworld, but many of them receive reinforcements as logistics allow. Cain’s regiment regularly returns to a staging world so they get a steady trickle, while the Ghosts are continuously on crusade forwards so their reinforcements from Belladon got stuck in transit for literally years, though admittedly that was de-prioritized because they were a band, and their Verghast reinforcements arrived as six formed companies.

          It seems like in most cases it’s just impractical to keep up a steady stream of reinforcements from the homeworlds, so they recruit replacements onsite or merge units and raise fresh ones as complete regiments from time to time.

    2. It’s obviously going to depend to an extent on which faction you’re talking about, but I feel like on the Imperial side, even if de-emphasised, the tooth-to-tail ratio is probably not far off the mark. Space Marines have massive chapter fortresses which dominate entire worlds, to mobilise and keep in the field a thousand fighting men (and noting that it’s rare that the whole chapter is ever in action simultaneously, with around a third of it acting largely as reserves), plus the chapter serfs which are implied to outnumber the Marines.

      The Imperial Guard has billions of frontline soldiers, but, even assuming the author in question ignores the question of logistics within the Guard itself, it also has the Mechanicum, which fills a lot of the support role that in modern armies is done “in-house”, and while the Mechanicum does have a fighting component (of course, because they’ve got to sell models somehow) its principal role is a support one, meaning that there is a whole, huge organisation standing behind the Guard as part of its “tail”, leaving aside any question of a “tail” within the Guard itself.

      The Tyranids strike me as another faction where the tooth:tail ratio is probably at least impliedly pretty solid, and while I don’t pay attention to the Tau I think their caste system impliedly allows for large water/earth-caste “tails” supporting the Fire Caste frontline troops?

      As ever, of course, it’s going to depend what sources you look at and how much the relevant author has thought about things. 40K remains eternally vulnerable to someone just coming along and rewriting whole swathes of background for the worse because they’re lazy or unimaginative.

      1. It will never amaze me how often 40k aims to be over-the-top, and in doing so ends up in the right ballpark for the scale of a million-world empire.

        Re: Tyranids: I’m pretty sure the tooth-to-tail ratio of a tyranid army is going to be in the triple digits—all tyranids have tons of teeth, but not all of them have tails.
        Jokes aside, their military biology is so alien to the kinds of armies that the tooth-to-tail ratio was designed to analyze that I don’t think you can calculate one. It’s bad enough when a society doesn’t have any full-time teeth to count; how much worse is it when you need to decide whether to count the soldiers’ weapons (often grown separately and then fused to the tyranid) as separate soldiers? Not to mention that it would just feel kinda wrong to count a gaunt and a norn-queen equally…

      2. Tyranid “tail” for ground forces purposes probably considerably outmasses tooth, when you consider the hive ships and capillary towers and digestion pools that do most of the logistics work. Plus their primary logistics work is foraging, which will occupy a decent chunk of their gaunts and rippers at any given time; they like to place enemy force concentrations under siege and go foraging to grow their heavy stuff on site.

        The space marines have either an absolute ton of chapter serfs, even discounting their ship crews, or a pile of servitors to do most of their logistics.

      3. For all that, there are about 1000 Space Marine chapters, in an empire of a million worlds. So they only really get about 0.1% of the resources dedicated to them, overall. Essentially, on galactic scale, they’re practically negligible compared to the Imperial Guard and Imperial Navy, which draw on resources of various forge worlds, hive worlds, orbitals, etc. which probably comes out to a sizable chunk of the resources in multiple percentage points of the entire Imperium dedicated to supporting, equipping and supplying them.

        1. all those million worlds are not created equally, like not all those worlds are hive or forge worlds, where most of the production is done, but rather producing primary resources or just low productivity/habitability.
          And I’d say the space marines tend to hog a good portion of the ‘high-tech’ (such as it can be called in the IoM) production, if only to have an explanation for the limited number of space marines.

      4. The Imperial Guard technically has two tails: those elements of the Mechanicus which are dedicated to assisting the Guard, and the Departmento Munitorium, which is a sub-department of the Imperium’s overall bureaucracy (the Administratum) specifically dedicated to organizing and overseeing the Guards logistics. In terms of portrayal, the Departmento Munitorium is essentially a miniature version of the Administratum: a MASSIVE harking bureaucracy whose sheer size leads to plenty of dysfunction yet it manages to keep plodding forward through it’s own immensity.

    3. Orkz would be all tooth (fitting), same with tyranids.
      The imperium suffers from the inefficiency of the lack of access and organization of the planets they control. They adopted the two tier army approach as outlined by mr. Devereaux. With the imperial guard protecting and holding most planets, and the space marines being the expensive high quality assault troops (the imperial guard also attack since the space marines are stretched too thin). And I think a lot of other races have similar low tooth to tail ratios. They all need to perform space travel, advanced military gear and advanced war machines all requires a lot of maintenance.

      1. It’s actually a three tier army. The overwhelming majority of people under arms in the IoM are “Planetary defense forces”, militas that are never expected to fight off their home planet and rarely too far from their home cities. The Imperial Guard does have much more of a widespread presence than the space marines, they’re usually called into a situation; most planets won’t have an IG garrison, but if they’re attacked they’ll scream for help from the closest IG force. IG are also used on offensive operations, where PDF are pretty much never doing that.

        1. It’s tough to classify gretchins as people, when the orks use them more like ambulatory rations.

        2. It’s really hard to determine where the tooth ends and the tail (or rations) begin with the orks, especially since they can just replenish their losses by waiting for more orks (or gretchin) to grow. Ork armies and ork logistics don’t work like they do for da hoomies, so using analytical techniques designed for human armies on them seems like a futile exercise.

          Ignoring that…since gretchin are a valid (and, given their point cost, surprisingly effective) choice of unit in an ork army, I’d probably put them in the same category as Roman combat engineers: “Tooth,” but with an asterisk.

      2. IIRC at least as of last time I paid much attention to the canon even the Imperial Guard were more on the “elite assault troops” end of the spectrum, with Planetary Defense Forces being the lower quality defensive units that hold space in rear areas.

      3. The Imperium has a dozen tiers for a dozen different political factions. The Astartes are high quality assault troops, but they’re under the purview of commanders who are politically independent from the military chain of command. The guard ranges drastically in quality from the Cadians/Catachans at one end, to the Salvar Chem Dogs/Tangar Woad Warriors at the other. Their tail is largely the Administratum, so is a seperate political entity. The other teeth that relies on the Administratum is the Navy, which of course is a Space Navy. Boats and stuff, when they come up, are the purview of specialised Guard regiments. The PDF answers to the planetary governor, although some sources show that the Guard can take command where expedient. The Inquisition has several separate armies at its disposal, let alone the personal paramilitaries some Inquisitors can raise. The Sisters of Battle are a millitary organisation linked to the Administratum, the Ecclesiarchy, the Inquisition, etc, but are really their own political faction. The Mechanicus is in some places/sources almost a separate allied empire, in others its a major Imperial faction – so obviously they have several militaries. Then there’s Knight Houses, etc. The Imperium really doesn’t have a cohesive anything, and their militaries are no exception. Each army operates under its own paradigm, and the resulting mess works very inefficiently, but nobody every said Fascist Space-Feudalism was efficient.

        1. And at least the Imperium is always faithfully portrayed as vastly inefficient.
          With the Rogue Trader RPG giving us concrete ship crew sizes, we can also say that Space Marine chapters are actually normal human navies (albeit consisting of very well trained and equipped crew), with a tiny elite marine special forces element attached. Crew numbers for a cruiser are about 95,000. Battleships are obviously much higher. And Marine chapters are usually portrayed as having several battle ships and cruisers, backed by a number of escort vessels. We might reasonably expect that the total crew complement of a poorly equipped chapters navy (assuming 1 battleship, 3 cruisers and 7 escorts) is about 650,000. Compared to the roughly 1000 marines. In a well equipped first founding chapter like the Dark Angels or the Ultramarines we’d expect more, and heavier ships so a ration of 1000 serfs to 1 marine seems about right. And that’s just the navy. It doesn’t include the huge numbers of serfs that maintain and run the fortresses, guard them, operate the defensive systems of the fortresses, etc.
          To be honest, the 1000 marines a chapter was always vastly too low.

          1. Legally speaking, space marine ships are for supporting planetary assaults and they are generally not permitted to have dedicated warships above frigate class, with notable exceptions in the form of the Phalanx and the surviving Glorianas. IIRC their personnel requirements are notably below average because they have significantly heavier use of automation, but it’s still much larger than the marines.

            Even discounting the crews, because the augmented guys hauling the torpedoes into their tubes aren’t doing ground logistics, there’s usually several serfs or arming servitors supporting each marine, and a good number more under the techmarines working on the vehicles. They’re generally strictly tail and won’t be deployed into combat; if they tried drop podding they’d break their spines.

          2. I’d note that the supporting planetary assaults rule isn’t just a fig leaf; their main guns tend to be “bombardment cannons” because they’re used for planetary bombardment, and they’ve all got extensive launch bays used for rapid gunship deployments. They’re effective space combatants because the gunships do double duty as attack craft and their marine complement can launch offensive boarding actions to devastating effect. They’re very high quality and their crews are excellent but they are undergunned for their size.

            This is also part of why imperial military matters are such a mess; Gulliman deliberately broke them so no single individual could control a significant portion of the imperial military again. Even among the High Lords of Terra command of the navy and ground forces is split.

    4. The Imperial Guard is probably somewhere around WWII levels. They’re mostly styled after WWII and WWI armies, they explicitly rely on manpower more than technology, and they aren’t a fully orbit-mobile force like other fictional space armies.

      They have some features that reduce their logistical burden – lasguns can be recharged by basically any source of energy, even a campfire, and Leman Russ tanks are noted to run on almost anything that burns. Their weapons are also noted to be extremely simple to operate, since they might end up being used by an illiterate conscript from a feral world who has never seen a gun in his life. They also don’t have guided weapons, night vision goggles, or any of the other fancy gear that modern soldiers get (except in small quantities for their elite units).

      On the other hand, their artillery and tanks are much bigger than their real-world counterparts (being modeled more after WWI than WWII). So while their infantry probably runs with very light supply lines, their armor and artillery are going to be hideously expensive to keep supplied. Which probably explains why the Guard seems to rely on mass infantry spam.

      (You also have to contend with the Administratum, which is so hideously bureaucratic that it’s possible for Guard regiments to die of old age while waiting for their orders to arrive. So like, take a reasonable estimate for a WWII army, and then quintuple it because you’ll be lucky if a fifth of your army even arrives on the right planet.)

      1. I have the distinct impression that Imperial war machines and tech in general are really rugged, and in earth-like environments don’t need to replace their tracks every hundred kilometers and suchlike. In harsher environments their logistics struggle a bit.

        1. Yep, a lot of them, while they LOOK like regular WWI machines, are actually very advanced, very rugged machines designed by a hyperintelligent AI (possibly, “What is an STC actually?” is a question) that are reproduced without anyone knowing how they *really* work.

        2. Chimeras and Rhinos, the standard transport vehicles, are built on a tractor chassis. A tractor chassis that was designed to be used by planetary colonization efforts, so it was designed knowing that it can’t be put into a shop every 100,000 kms for an overhaul. That’s a machine that was intended to run for decades or even generations without anything more complex than an oil change.

          1. Still needs shells, though. Between the 3 heavy bolters and the main cannon, a Leman Russ will need a lot of ammo.

            (I also thought the shells would be considerably bigger and more cumbersome, since it’s usually depicted with a gun big enough to fit a small child, but apparently official sources say it’s a 120mm cannon, similar to modern MBTs. Surprisingly reasonable.)

    5. The armies are ridiculously large but then so is the munitorium , the military bureaucracy. In the novels one only meets the tooth by some include camp followers and practically all complain about the munitorium.

      In the grimdark of the far future there is only war. And bureaucracy.

      1. I remember one comment (but can’t point at the source) how the Administratum branch that deal with the Imperial military (Departmento Munitorum) has one pencil-pusher per soldier in the active regiments, so that’d give a 1:1 tooth to tail ratio, discounting how the regiments have organic tail components: I think in one of the Cain books, one of the five company of the regiment he’s attached to is explicitly composed of ‘tail’ soldiers.

    6. I always like the 40k depiction of its tail, since it’s always from the marines’ point of view and it’s “It’s a terrible loss as two of our battle brothers fell oh and by the way a hundred thousand non-combatants also got jettisoned out into the void of space, but we will always remember Battle Brothers Marius and Hruthgar.”

    7. Despite all the comments on its immense size, most Imperial Guard armies are in the high hundreds of thousands and low millions. Major planetary operations are carried out with half a million men, large theatres account for 10 to 20 million troops. Of course, there are exceptions, such as Crusades which mobilise huge fractions of the Imperium/ resources – the Sabbat Worlds Crusade had a billion Guardsmen, though this is on the high hens for Crusades, most seem to be closer to a hundred million troops.
      Given all this, it’s no wonder the Departmento Munitorum and the Administratum are so massive, they probably outmass the Guard several times over

    8. It occurs to me that the tooth to tail ratio is lower than it might otherwise be because the tail is heavily automated in the form of servitors. They’re often briefly mentioned performing menial maintanence or transporting heavy loads and may perform the bulk of the physical labor with humans supervising. They’d be classed as equipment rather than personnel, and in the Great Crusade era they’re common enough that Olli got to take one with him when he retired from being a line trooper, and a relatively sophisticated one at that.

      It’s one of the casually brushed-over horrors of the Imperium, that while many servitors are vat-grown the bulk are criminals who had their minds wiped, and none of our viewpoint characters even blink. Given the sheer number of them you have to wonder how many of them were arrested for stuff like not paying their debts because they lost their jobs.

  2. I kinda disagree with the argument about smaller fleets because ‘spaceships are big’. Yes, spaceships are big but industrial output of a proper interstellar civilization is also enormous.

    Fully colonized star akin to our Sun will probably have capacity to build Star Destroyer equivalents by the millions. And of course Star Destroyer is a relatively small for the interstellar ship. Black hole-based propulsion/power-system alone would have mass in the millions of tons, counting an artificial black hole itself, a laser array to maintain and feed it and required heat management systems for all of that.

    1. You’re assuming a K2-scale civilization, not a typical-space-opera-empire-scale civilization. The former has potentially quintillions of inhabitants and the industrial capacity that implies; the latter has thousands or millions of planets, most inhabited by a large city’s worth of people or less, for a total population around the trillions. The difference between the former and the latter (in terms of population, industrial capacity, whatever) is roughly equivalent to the difference between the entire modern Earth and a very small town.

    2. I agree with GreatWyrmGold, there’s a big difference between the theoretical capacity of a star system and the capacity the inhabitants can use. In e.g. Star Wars the technology available is much closer to us than to a civilization that creates artificial black holes as power plants and the average *important* planet seems to have only a few cities of note with a few ultra-dense outliers.

      TLDR The Empire can’t build a million super-ships with one star system for the same reason we can’t.

    3. Most SciFi reads more like K1-scale civilizations than K2-scales. Also, the square-cube law makes star-destroyer scale spacecraft utterly impractical, especially for actual combat and maneuver, barring new physics.

      1. How does the square-cube law do that? Why would a mile-long spaceship be any more impossible than a mile-high building? We could surely build something like that right now, if anyone had good reason to.

        1. Heat is one problem that most SF ignores. Heat generation goes with volume, usually cube; heat radiation goes with area, square. And in the vacuum of space, radiation is your only sustainable way to get rid of heat.

          OTOH radiation shielding might favor large motherships; if you need a few meters of rock to be safe, that’s less of a design ‘tax’ if your ship is 100s of meters across rather than 10.

          Then there’s gravity. If you’re using rotation for gravity, there’s a sweet range between “too small and you’re dizzy” and “too big and you rip yourself apart”. That’s not a cube-square problem, though.

        2. Last I heard, we can’t build mile-high buildings because there is no current building material that can support its own weight if you stack it a mile high.

          1. Actually mile-high towers are technically possible within the strength/weight limits of structural steel and have been proposed in the past. They simply aren’t practical due to cost and logistics issues.

    4. This is also what I think. If you are building multiple star destroyers, it makes sense to start building them in orbit already (so you don’t have to make sure each intermediate building stage can handle gravity). Then it makes sense to source materials from asteroids. Then it makes sense to start building big outposts on or near asteroids. And quite quickly you will have much ore ability to build and supply starships without the 10x the logistics cost that you would have if you had to lift everything into orbit first.

      At that point, each system should be able to have its own starship factories making full star destroyers with attendant ships. But then you are also not in the star wars universe anymore.

      1. On Star Destroyers, you must bear in mind the design philosophy as well. The Imperial navy didn’t actually have any competing power that could really build warships in anything like the same scale. They had no military peers in that sense. Instead what they went for was the Tarkin principle. better to rule through fear of force than through force alone. So star destroyers are larger than they really need to be and scary looking. They are true multi-role ships being carriers, battleships and also containing everything you need to launch ground operations. A squadron of star destroyers jumping into orbit above a rebellious planet was a terrifying sign to stop mucking around and behave. They were never really designed to be front line war fighting machines, because there was no possible force to match the Imperial navy. Even the Rebel Alliance did everything they could to avoid pitched battles with them.
        And they built them at scale. The Imperial navy had about 25,000 star destroyers. Vastly more capital ships than anyone else. But they also had like half a galaxy to patrol. So their fleets were often very spread out.

  3. Tooth to tail ratios will also grow as you get into galaxy-spanning science fiction/space opera for the same reason that large structures don’t grow linearly. Think about an airplane wing. It has to support all of the structure outboard of the current location of the wing. So as the wing gets longer you have to add more and more inboard structure to support the weight of that outboard structure. It doesn’t grow linearly.

    Logistic systems have support the logistic systems that get added on to support the distance (and complexity!) being supported. So to make a major effort against a faraway planet you have to have all of the self-contained logistics to function at the planet, and then all of the logistical systems to support getting out to that faraway planet in the first place. Those systems grow exponentially, not linearly.

    For those who are familiar with the “tyranny of the rocket equation”, it’s a similar thing: As you scale up distance you have to carry the fuel to transport the fuel that gets you a little bit farther out.

    To support a single one of Robert Heinlein’s Marauder suits out at the sharp end of the stick would require a massive logistical system. I can conceive of tooth-to-tail ratios of 1:100 or 1:1000 pretty easily.

    1. Explains why the MI are in such low numbers and have to rely on the massive force multipliers.

      1. Heinlein also touches upon a related matter: while the MI is incredibly dependent upon a massive amount of logistical and other support, the individuals providing that support are not members of the MI, nor do they share an equivalent rank structure.

        Contrast the modern American ground forces, which have an almost Glasgow football match level of tension between “grunts” and “POGS.”

        1. And even with all that Navy support, their usual operation pattern is drop, fight for a few hours at most, and then return to the ship. Where they then spend the next several weeks doing maintenance on their powered armor suits.

  4. Pedantically jumping into Dune, I would point out a couple of other factors: 1, the Spacing Guild does not like it when armies move around, the limiting factor of that maximum Harkonen effort involves the expense of moving troops into combat zones, not anything about raising them or equipping them. And since they have a monopoly on space travel, it’s their way or the highway. (Timothy Zahn’s Quadrail books take a similar take on this sort of thing, although the underlying structure is a bit different)

    I also wonder about the campaigning on Arrakis. As far as the Imperium is concerned, Arrakis has almost all of its population concentrated in a smallish area around the north pole of the planet. I found a reasonably good map here but take a look at where those villages are, they’re all clustered in one smallish part of the map. Especially in sci-fi settings where people settle on all sorts of inhospitable rocks, “control of the planet” might mean only needing to control an area the size of a small country here on Earth.

    I also wonder about, and pardon me if I’m imprecise, as I don’t know a definitive term for the concept and I’m groping towards it:, but about what I’m going to call a soldier’s field of control. Back when Putin was building up but hadn’t attacked Ukraine yet, the dominant narrative I was hearing in articles and on the news was that Putin could probably overrun Ukraine pretty quickly (oops) but that actually controlling Ukraine in the face of an insurgency that was likely to develop was far more doubtful. At the end of the day, the force he had amassed was too small to stand over every needed street corner and watch things and keep the conquests secure. It’s very easy to imagine futuristic technology making an individual soldier have far more firepower than exists here on earth. But if someone wants to conquer a planet, presumably they also want to hold that planet and reorder things about it. And that probably means occupying a lot of space and pacifying a lot of potentially hostile civilians.

    In real life here on Earth, that’s usually accomplished by big occupation forces trying to keep eyes on a lot of places and ready to respond if trouble breaks out at any number of points. I’m wondering how futuristic technology can alter that balance. Would it give insurgents more places to hide or more ways of masking their activities? Would it allow occupying soldiers to see more at once or direct fires from further away and more accurately? How does futuristic communication affect organization or give possibility of raising awareness of an occupied planet to a wider audience?

    1. “Especially in sci-fi settings where people settle on all sorts of inhospitable rocks, “control of the planet” might mean only needing to control an area the size of a small country here on Earth.”

      I had the same thought with regards to Arrakis… It’s not the industrial era model where you all but have to physically conquer and occupy broad swathes of territory in order to subdue it. It’s more the older model, where the fall of single city, a single castle can lead to wildly disproportionate outcomes in terms of influence and control.

      Control Arrakeen and you have effective control of the planet.

      This too falls directly out of faufreluches – the keystone of the system is the individual. And the system does not survive the death of the individual (and his heirs).

    2. To answer your second part: It depends a lot on both what futuristic technology you’re assuming and how effectively the invaders can take control of existing technological infrastructure. For instance, if we’re assuming the planetary government uses AI surveillance and robo-cops, occupation would require almost no manpower if the invaders could subvert those systems and be virtually impossible if the police AI cooperated with the insurgents.

    3. I would imagine drones and automated recognition systems, informing quick-response tactical units, would cover the vast majority of surveillance and patrol requirements.

      1. Dune explicitly has no automation.

        “Thou shalt not make a machine in the likeness of a human mind.”

        1. Dune most definitely has automation. It doesn’t have full AI, but things like the Hunter-Seeker or the sparring bot Alia goes up against in Children of Dune are definitely machines that can act with a degree of self-direction.

          1. I don’t remember details like that, but I do recall that Ix has more advanced machines than the rest of the milieu, so if either of those were from Ix then they don’t indicate what’s normally available.

          2. The Hunter-Seeker flying needle in the first Dune book is remote controlled by a human operator, no onboard autopilot.

            The sparring bot in Children of Dune is owned by a member of the royal family and kept in her private quarters. It’s the only vaguely automated machine in the first three books, and doesn’t seem to do any more than one of our own 20th C industrial robot arms could do.

            God Emperor of Dune has a high ranking member of the Emperor’s personal staff using a computer, which is still described as “forbidden”.

            My guess is that royalty and very important people can get away with the occasional bit of automation as long as they don’t flaunt it.

            By the time of Chapter House of Dune the prohibition is finally gone, there are computers again. At least three thousand years after Dune though.

          3. I’m not sure which FH Dune book it is in, but the Bene Gesserit always had covert computers. They didn’t run the breeding program with index cards and filing cabinets.

    4. One possible scenario in space-opera warfare would be for major interstellar powers to lean into the “high-low mix” concept. The Galactic Empire has a comparatively small force of highly mobile, heavily armed troops who use the latest in advanced communications, coordinated bombardment, and expensive but powerful equipment to sweep enemies away in field battles.

      Then they also have the COIN troopers.

      These are the guys whose equipment consists of standard light infantry fare. Nothing designed to need extensive and constant maintenance. Ideally, very little that can’t be maintained on-planet using the available resources and facilities of a world you’ve already conquered. They have almost no heavy armor beyond what is necessary to protect against the typical weapons available to guerilla fighters. Their artillery assets are quite limited, and possibly they rely on bombardment platforms parked in orbit by the Imperial Navy and likewise designed to operate for extended periods with limited maintenance and resupply.

      The COIN troopers would get slaughtered if fighting a heavily equipped force that had all the supply and firepower advantages of (post-)industrial warfare… But they’re not intended to. They’re intended to serve as a force that can at least plausibly be recruited, trained, and deployed in large enough numbers to occupy entire planetary populations with “a trooper on every street corner,” while keeping tooth-to-tail ratios to a manageable level that permits this.

    5. It is also, if I remember, a plot point that the Harkonnens have massively underestimated the actual population of Arrakis, which is why they have such a hard time trying to hold on to it. After the invasion, the Baron says explicitly that he estimates the population at five million – at a standard counterinsurgency ratio of 50:1, ten legions should be enough to hold that down. He’s shocked to hear Thufir Hawat tell him that, no, the number is more like 15 million. (That should still have been just within the power of a force of 300,000 troops, but Fremen are terribly tough, and he may well have withdrawn a lot of the invasion force now that the big MISSION ACCOMPLISHED banner has gone up.)

  5. I’ve been enjoying Lois McMaster Bujold’s Vorkosigan Saga for a few months now. It makes me think of this blog because the protagonists spend a lot more time figuring out how to get paid or keep supplies up than they do in actual combat.

    1. Based on that alone it sounds like Lois McMaster Bujold has a decent understanding of how most militaries work.

  6. The immunes included engineers, catapult-operators, musicians—

    Wait, what was that last one? Does Devereux just mean trumpeteers and drummers and such for communication/coordination, or did the Romans pay pipers to…um…morale or something?

      1. Musicians were also used by athletes in training. No stopwatches, so find a musician and count beats.
        (Courtesy of the Ancient Greeks exhibition from the British Museum which is currently on display in my city.)

    1. British Napoleonic armies had musicians, notably scots regiments whose pipers would play as they marched into battle. The musicians doubled as medics.

      1. They still have them, and the musicians are generally still medics (not always). The last time that a Highland regiment marched into battle with pipes playing was, depending on your definition of “battle”, either 1944, 1965, or 2004.

    2. Both. Just because Marius Q Centurion is paid to play his pipes for military formations doesn’t mean he can’t play a fun little ditty in his off hours to keep morale up.

  7. “ships in low orbit circle a planet very fast, enabling them to deorbit basically anywhere on very short notice”

    That capability is… often overstated. Orbits and re-entry trajectories are ballistic, and for a modern civilization trivially predictable. Orbital inclination also matters, a spaceship in 30 degree orbit around the Earth isn’t going to suddenly drop down on Moscow (which it 55 degrees North)… Even if it is at the proper position to initiate retrofire. (Which it will only be *once* in a particular orbit, and that particular orbit only occurs at longer intervals as the planet rotates under the orbital plane.)

    For realistic spacecraft, orbital mechanics are just as tyrannical as the rocket equation.

    1. If you have ships in orbit around an enemy planet, you don’t have realistic spacecraft.

    2. In a polar orbit, even with minimal cross-range capability, you can drop anywhere on a plenty with at most one days’ notice.

  8. “Whole sections of the capital planet seem to exist effectively outside of the supervision of the central government”

    Modern administrative states on Earth also don’t have full control over their capitals. The neglected areas are usually called ghettos, flavelas, no-go zones, etc.

    1. Coruscant is significantly worse. At least in some of the fiction, they basically have neither visibility nor presence in massive parts of the world (which makes sense, as otherwise Coruscant ought to be able to conquer most of the rest of the galaxy given the implied population and industrial capacity of a world-city).

      1. Coruscant is, however, poorly thought out. A world-city could only exist by pulling enormous resources (food, raw materials, etc) from other worlds to support itself. This is stated to be the case with Asimov’s Trantor, but Trantor governs a much larger, and more importantly, much more centralized empire. It has a larger administrative role to fill and a non-democratic system that can demand those resources from those it rules over. The Republic (under which Coruscant was built up) is much more centralized (meaning it doesn’t need the same administrative capacity) and democratic. (So people can refuse to hand over those resources) There seems to be little reasons for Coruscant to be so over-developed, except that Lucas had probably read Asimov and wanted to imitate Trantor.

        1. As far as I can tell, the Star Wars Republic doesn’t even have a military prior to the clone army. It seems to resemble the European Union, or even the United Nations, more than it resembles a state. Its members have to uphold certain moral standards (no flesh-and-blood slaves), but are mostly sovereign.

          As for how Coruscant gets fed, I don’t know if there’s any official explanation. But, extending the U.N. metaphor further, how New York City feeds itself has little to do with it being the capital of the U.N.

          1. Not much of a government. It does have the Jedi Order though, which functions as a quasi-military institution, plus it has a least some ships. (Like whatever it is the brings Obi-Wan and Qui-Gon to the Trade Federation ship in E1) In a short piece I wrote, I argued that it is essential an confederation, like the HRE or the US under the Articles of Confederation. A government, but a radically decentralized and weak one.

            As for feeding and otherwise supplying itself, a planet city isn’t going to produce enough of its own food (barring an unseen technological fix), and presumably it doesn’t mine enough resources to feed at ton of industries. Therefore, you have to import everything, and do so by star ship, which would be costly on that scale. I would argue that there is little solid economic reason for such a planet to exist. It does so because it serves another purpose, such as housing the administration necessary to run a galactic state. But my original point was that Coruscant doesn’t do that to the same degree that Trantor did.

          2. In Star Wars it is economically viable to transport raw resources to other planets for manufacturing. Coruscant isn’t the only city planet and there’s major shipyard worlds that rely on mining worlds for import. Yes, that requires highly cost-effective interstellar transport, which Star Wars has.

        2. Food?

          How about oxygen?

          No city-planet is going to manage without LARGE parks to do the oxygen exchange. And then you hit the problem that decaying plant matter consumes as much oxygen as it generated, so you have to either raise food with as much of the plant being eaten as possible (salad, everyone?) so that the oxygen breathing does the break down, or fix the carbon as permanently as possible (wood furnishings? pump it somewhere to turn into coal or something?).

          1. Not to mention that a net importation of food means a net exportation of sewage.

          2. I’ve also read that the body heat from all of those people would be unworkable. But Star Wars physics is so bonkers I don’t even worry about it.

          3. Harry Harrison’s “Bill The Galactic Hero” addresses the oxygen problem (and the trash disposal problem) in its parody of Trantor. Besides the food ships there are ships bringing in oxygen and hauling away CO2.

          4. Star Wars operates on a pretty big scale and has a lot of soft-SF technology that “shouldn’t work” in real life (force fields and such are obvious examples but not the only one).

            It’s at least conceivable that Coruscant is relying on artificial machinery to circulate and “scrub” the atmosphere of CO2, as opposed to photosynthesis from plants.

        3. Tarkin refers to the “million systems of the galactic empire”. It’s pretty big.

          As for centralization, I don’t think very much of Coruscant is there to support the administrative apparatus, it’s there to be a city and do all the things cities do; lots of cities these days aren’t capitals at all and exert no administrative control over the areas that provide them with food or raw materials.

          1. I don’t know that Tarkin quote; can you find a clip of it? In Ep. 2, thousands of systems are joining the Separatists, which is enough to split the Republic, so I’d guess more like 10,000 inhabited systems, going off of the films.

          2. Novelization of A New Hope.

            If working just from the films, I’d take the position that the Empire/Republic is large enough to support Coruscant. Which does lead to the question of how the revolt of thousands of systems is a serious threat even given that the pre-clone Republic, as far as we can tell, has a military comprised of the Jedi and the Senate guard, but they seem to be key economic powers like the Trade Federation and probably control a disproportionate share of industrial power.

          3. The line for Episode II: “As I explained to you earlier, I am quite convinced that 10,000 more systems will rally to our cause with your support, gentlemen.”

            I’d like to draw attention to the word “more”.

            Also, not all systems are equal. I think it’s possible that Tarkin is including systems with extremely small populations that most people would ignore.

        4. Trantor isn’t well-thought out either. Supposedly the entire land area is covered in multi-story arcology, yet it has only 40 billion people, just 5x the current population of the Earth. At Manhattan densities, 40 billion people just need 1.6 million km2 — California, Arizona, New Mexico, and Texas combined.

          As for Coruscant… well, depends on the tech being applied. Humans are roughly 100 Watt heat lamps. If we assume some future tech can take 1000 W of power and synthesize 100 W of tasty food, then 2e12 organic sapients need 2e15 W of power. If solar power farms can get 100 W/m2 on average, that’s 2e13 m2, or 20 million km2. Earth has a land area of around 150 million km2. So say 80 million km2 for them to live in at Manhattan densities, 60 million km2 for solar power plants (for power beyond just food), plus managing the oceans to produce more stuff, and you can kind of make it work without any major imports.

          (Me, I always thought it was supposed to be 1 trillion people, but Wikipedia says 2 and Wookiepedia says “trillions”. It may have crept. But with the right tech you could certainly have hundreds of billions of people.)

          You could also postulate non-solar power sources, but then we start having to worry about the planet’s power budget, i.e. cooking the planet from too many nuclear power plants…

          1. You’re right about Trantor’s population. I wouldn’t argue it makes sense in all respects, just that there is a better reason for a world-city to exist there than on Coruscant. It’s not that you couldn’t conceivably make Coruscant work, it’s that there’s no good reason in setting to do so.

          2. I’ve only seen movies 4-6 and 1, and suspect we saw more of Coruscant in 2&3. But is there strong canon reason to rule out it being an economic center, like NYC? Wookiepedia says it was at the “end” of several trade routes.

          3. Star Wars’ extended universes (Disney did a lore reset), both go with the trade hub justification. Both EUs (mostly) make it much easier to travel along ‘hyperlanes” and put Coruscant at the only intersection of the two most important ones (the Corellian Run and Perlemian). Also, Coruscant is (probably) humanities home planet in Star Wars canon so it gets a lot of the boost that Earth gets in other franchises.

          4. One of my favorite bits of math is that a population of little over 215.7 billion people on Earth would give us the average population density of… the Netherlands. That famously overbuilt and city-dense nation.

            That last sentence is sarcasm, to be clear.

          5. >Trantor isn’t well-thought out either.
            >Supposedly the entire land area is
            >covered in multi-story arcology, yet
            >it has only 40 billion people, just 5x
            >the current population of the Earth.
            >At Manhattan densities, 40 billion
            >people just need 1.6 million km2
            >— California, Arizona, New Mexico,
            >and Texas combined.

            The actual first description of Trantor as a city-world in Asimov’s writing is: “All the land surface of Trantor, 75,000,000 square miles in extent, was a single city. The population, at its height, was well in excess of forty billions.” And the following text does make it clear that the normal lifestyle of a Trantorian during the Galactic Empire was to spend effectively one’s whole life under some sort of artificial roof, and that nearly the whole planet is covered in metal.

            So yeah, I don’t think Asimov really sat down and did the math. And he was assuredly capable of doing so- went on to get a doctorate in biochemistry.

            In fairness to him, 40-45 billion was a stunningly large population to imagine from the perspective of 1940, because that was before the Green Revolution, when it was generally believed that the agricultural carrying capacity of the Earth was… Well, let’s just say “a lot less than the real life population of the Earth today, in 2022.”

        5. It can pull resources to itself via the free market, no state power required. So that part isn’t the poorly thought out part.

          1. Yes, but my argument is that it wouldn’t grow to that size via the free market, because there wouldn’t be a compelling economic reason for it to do so. Corsucant wouldn’t be near enough natural resources to provide a comparative advantage in that regard and labor costs would be much higher than on a less populated world. Food would have to be imported, waste off-loaded, housing costs would be high, necessitating higher wages, etc. (The hyperspace routes argument mentioned above could help a bit, but comes from EU sources, is not supported by the films, and still wouldn’t explain that much growth)

            Of course, this is true of most cities, esp. the higher labor costs. They grow in spite of that because they’re centers for administration, both governmental and economic, which forces lots of jobs to move there, followed by secondary jobs to support those workers. (restaurants, stores, etc)

            But this has its limits. No global city has a population of even 40m for its extended urban area. As someone mentioned, we’re probably talking 1 trillion+ for Coruscant as depicted. The only way it would make sense for it to get that big is if the administration it supported was comparably enormous. But as also discussed above, the Republic is super decentralized. It has a minimal military, very weak central gov, etc. Therefore, I find it doubtful that would required a planet city to support it. (Unlike Trantor, which supports a much larger empire and a more centralized one)

          2. Cheap easy gravity control does make travel up and down five-kilometer towers more practical than it would be in real life.

          3. paulewenstein, as I understand it, London has had several % of the population of the British Isles, entirely through free-market forces, since before the Industrial Revolution. How big a population would Coruscant have if it had several % of the population of the Galactic Empire?

          4. replying to ad (below) and Lekhaka – anyone who thinks either London or the Netherlands garnered their populations and command of resources through the free market does not know any economic history.

          5. It’s quite possible that the Trade Federation (to pick one known Star Wars megacorp) has huge numbers of employees working on Coruscant. Not just the obvious populations of lobbyists and whatnot, but that this is where you keep a lot of the big corporate offices, the stock exchanges, and so on. Because it’s the center; it’s where you go for prestige and to network with other powerful organizations. Multiply this by all the other big interstellar corporations, and in practice there may still be billions upon billions of administrative workers living on Coruscant and making decisions that shape the galaxy… It’s just that most of those workers aren’t public service sector employees.

            Also, Coruscant has been the capital of the Republic for many millennia. Even if, as of the movie era, the Republic’s government is decentralized, it may have been more centralized in the past. Maybe Coruscant became a city-world ten thousand years ago when the Republic had huge combat fleets and built its naval shipyards on Coruscant. Maybe it first attracted hundreds of billions of bureaucrats during a period of exceptional centralization three thousand years ago, and the rest of the planetary economy grew up around that, and even when the central government laid off a lot of the bureaucrats in subsequent centuries, the economy just kept running on inertia.

  9. I feel like a mecha series could do some interesting stuff with this idea of tooth-to-tail ratios, since a Super Fighting Robot is sort of the ultimate reduction in teeth – a unit piloted by one person that’s often depicted taking on whole armies of previous-generation units. Extrapolate that trend out far enough and you have an empire whose “army” consists of a handful of Super Fighting Robots and their supply line, plus some second-tier units who mostly exist to occupy territory taken by the giant robots.

    1. An extreme example is single Evangelion unit requires an entire city’s worth of tail for a single soldier… XD

    2. The problem there, as wet navies have discovered, is that One Big Unit can only be in one place at one time. It’s radius of influence is limited by how fast it can find out about events of interest and how fast it can subsequently arrive on scene. That leaves gaps where smaller units can flourish… Or where the Other Guys One Big Unit can perform it’s dirty deeds and escape unmolested long before they can be effectively intercepted.

    3. Battletech actually is this in the fluff. Planetary invasions might involve maybe three dozen mechs invading a world guarded by the same. (Usually trying to capture the planetary capital) with maybe a few regiments of infantry and lightweight tanks (and a bit or artillery, some aerosoace fighters, etc.) In support of each. Really important worlds maybe get a regiment of 108 mechs. The players often complain about the small scale of the individual invasions but the setting is (with a few exceptions) supposed to be in times of restricted economic and industrial conditions, and the mechs ridiculously high powered. Which just sometimes clashes with the game rules and their gamebalance driven short ranges and such, and the grand scale of the settled worlds of the inner sphere. (That previous eras under a unified interstellar star league could involve throwing multiple thousand plus mech strong divisions around at single planets is often broght up for example.. but that was before four massive interstellar wars nuked most of the factories, universities, etc, and those massive armies blasted each other to scrap. Along with much of the population.)

      Later timeperiods increased the scale as technology and industry recovered some, but mostly just to the point of using regiments to conquers worlds and two or three for important ones, mostly because it makes for grander stories

      1. Yeah Battletech is my jam and I’ve talked about it at length for years, particularly with a friend who did a fair chunk of logistical support work in the Navy. We’ve talked about some of the ridiculous hammerspace concepts that involve fielding a battalion out of a single dropship that accounts for mechs and pilots but nowhere near the amount of space you’d need for parts, ammo, basic lubricants and other consumables, support personnel, etc. OG Battletech and a lot of the modern material just pretends that the tail barely exists, beyond a handful of mech technicians.

        The concept of the original game was basically feudalism in space, and the setting was originally built around that junkyard wars concept where nobles traded border worlds more out of hunger for glory and personal vendettas than any sort of greater nationalistic sentiment.

        In that reading, the border worlds generally accepted the occasional change in management because it didn’t greatly affect the lives of day to day citizens. Every couple of years, some rich jackasses would land a dropship and shoot it out with the current rich jackasses in charge. They’d sort their business out away from the cities and industrial centers, and occasionally the flag would change over the government buildings.

        That reading kind of works if you can accept a certain amount of handwaving (the amount of repair parts and supplies and so on that could be fit into a single dropship, and bearing in mind that the original sourcebooks had permanently damaged mechs still in service because nobody could afford to throw even the beaters away.)

        But then when they scaled the setting up for bigger, more dramatic stories they never really filled in the space it would take to depict actual planetary conquest, the sort of ongoing support such conquests would need, and so on. Battletech logistics are easily an order of magnitude away from where they’d need to be to account for the kind of grand campaigns they like to have now.

        Still a fun game, and easily adapted to account for more pragmatic readings. I’ve been running a campaign in it for years now, including a fairly involved storyline involving a rebellion being suppressed. The invading force (which the players were part of) ran into serious logistical issues early on because their oh-so-brilliant field marshal had assumed that they were going to win in three weeks. Six weeks in they’re scrambling to get the merchant marine network in place to support the invasion because the rather inconsiderate rebels hadn’t accommodated them by folding after taking a couple of good thrashings in open engagements.

        I have a policy in games (and in fiction in general)- realism is a seasoning, spice the setting to flavor, but don’t bog it down to the point where it kills the story.

  10. One thing that I like about both Halo and Star Wars is that they take seriously(ish) the ability to destroy planets via orbital bombardments.

    In Halo, planets basically can’t be defended from this, and so armies don’t really seem to exist: Master Chief et al are “marines” in the Napoleonic Wars sense of “dudes that defend ships from boarding parties”. Every weapon he has makes sense in this context: a “sniper” rifle that has about the same functional range as a Garand, a “main” “battle” “tank” that doesn’t have AP rounds and can be destroyed by small arms, etc.

    In Star Wars, shields allow planets to be defended from orbital bombardment, and so all the land battles we see are attempts at coups-de-main to shut down planetary defenses. There isn’t really any reason to put more boots on the ground than are necessary to accomplish that, hence the small size of the GAR, etc.

    1. 40k also has orbital bombardment in it, but it’s got several limiting factors.

      1. It’s not very accurate without a beacon on the ground; if you have a gun that’ll destroy a bunker only with a direct hit you’ll be here all day.
      2. There are void shields around a lot of truly major targets. The most spectacular example is the Imperial Palace shields that withstand bombardment by thousands of ships for literally months even with Perturbo calculating a firing pattern that causes some sort of interference with its recharge cycle.
      3. In a significant percentage of cases, battles are over something, be it raw minerals, arable land, industrial complexes, or the population. If you want to take the Urdeshi forges you can’t do that with orbital bombardment.
      4. In a small percentage of cases your target is not fully material and will not in fact die or even go away if you hit it in the face with a teraton-range nuke and must instead be stabbed with a magic knife.

      Thus a majority of the time orbital bombardment merely paves the way for landing mass forces to take the strategic targets. The Imperium pretty much only glasses planets if they can’t be taken by any military force that could plausibly be assembled or if the Tyranids are going to eat it.

  11. For tooth to tail it is plausible that the tail will be significantly more automated than the tooth either for technological or ethical reasons. That would decrease the relative number of people there.

    Firepower may well continue to mandate increasing dispersion, but the front changing from a 1 dimensional line two a 2 dimensional surface also potentially radically increases the effective size of the front you must cover.

    1. There’s also the philosophical question about when does does something have sufficient sentience to count toward the tooth or tail vs. just being a machine that gets omitted. Does R2-D2 in Luke’s X-Wing count toward the tooth? What about the droids in the prequels that are there in crazy numbers? Conversely, we see sentient droids repairing equipment and piloting logistics craft. Do they count toward the tail? I’m short, does something have to be organic to count toward tooth or tail?

      1. Bring up R2 makes me think of a real-life question: Do you have to pull a trigger to count as tooth, or is it enough to be a sailor on a vessel engaged in combat?

        1. Even within the ship, you have both tooth and tail, so you would look at the ratio within the ship. Sure, on a carrier you have pilots and air control officers…but you also have the mechanics to fix and rearm the planes, engineers to make the ship go where it’s supposed to and fix the damage that the pilots inevitably cause to their planes….and then you have to support both the teeth (the pilots) and the tail (everyone else) with even more tail, for which you have the cooks and barbers and doctors and quartermasters and counselors and tailors and launderers and gardeners…. Even on a non-carrier vessel, there are only so many guys manning the missiles or steering the ship.

          As a result, the “tooth-to-tail” ratio of a CBG is going to be pretty heavily slanted towards the tail

        2. Heuristic for tooth vs tail: Shooting and being shot at is an expected part of the job for tooth.
          Expanding, if the javelins / arrows / bullets / missiles are flying in both directions and you are meant to be there, you are tooth. If your own troops would be yelling “What the f**k are you doing here? Run away!” you are tail.
          This isn’t to say that tail don’t get shot at, they frequently do. And some tail troops carry weapons to defend themselves, at least in theory. But it is not part of The Plan for their own commanders.
          Plenty of military aircraft have crew who don’t pull a trigger, very often including the pilot. They are still tooth, the plane cannot fight without them.
          Warships have a large number of people who run the engines or haul ammo, but they don’t get offloaded before a battle. They are tooth. The crew on the tankers and replenishment ships may get shot at and even sunk by enemy action, but their own commanders work hard to keep this from happening so they can concentrate on logistics. They are tail.

      2. Sentient beings *should* count as being tooth or tail, so R2-D2 flying alongside Luke in an X-Wing is tooth, while repair and logistics droids would be tail.

        In the Star Wars universe droids are treated more as a slave underclass, like Spartan helots or worse. So while I would count them as part of the tooth/tail, I suspect in universe they get counted as equipment / supplies.

        1. Still, one can differentiate between a rifle and the wrench used to take off the tire while replacing it on the truck.

  12. Where do contractors fit into the tooth to tail ratio? In the US military today a lot of functions that used to be done by people in uniform are now done by civilians. A lot of bases / facilities use DoD Police instead of MPs. Janitorial and base up keep are civilian jobs. As are many mechanics positions – especially at the depots.

    I think this translates to a much larger tail then is at first obvious.

    1. It is by function, not formal membership of the military.
      Mercenaries are civilian rather than military but those who fight are still tooth. The Russian Wagner group is a current example, Blackwater in Iraq previously. Both have tooth and tail roles.
      The US and other western militaries do have a lot of contractors in the tail doing jobs that were formerly done by people who were officially part of the military. They should be counted, although yes probably some lazy writers / analysts don’t.
      Then again this isn’t some 21st C development. Much of the “tail” support functions for medieval and early modern European militaries, especially in navies, was contracted out.

      1. Blackwater is interesting, though, in that their primary job wasn’t to destroy the enemy. Their best-known contracts were providing security for officials and installations, especially diplomatic security. That’s a greater focus on lethality than, say, a mechanic. Yet using a commenter’s standard above that the difference between tooth and tail is whether you are meant to be in combat, I’d say clearly their main contracts were meant to avoid combat as much as possible. A good comparison would be to chaplain assistants — who in addition to helping with clergy tasks also are the defense for chaplains because the chaplains are non-combatants. By contrast, Wagner, appears to be used more as a traditional combat arms unit. In both Syria and Ukraine, the intent was for them to fight the enemy, although I believe I’ve read that they also provide logistics functions, as does Blackwater. (None of this is meant as a defense of Blackwater.)

  13. The former, except instead of drums they used a brass instrument called a cornu (Latin for an animal’s horn, which it resembled.)

    1. This was meant as a reply to the comment about musicians above. Someday I’ll stop hitting the wrong reply button, but clearly not today.

  14. > That said, an army structured for orbit-to-ground combat operations is going to have wild logistics demands, even if we assume lots of exotic technology solving things like, for instance, the prohibitive cost of getting military equipment to and from orbit.

    Isn’t this predicated on the assumption that things like fuel and ammunition are still significant logistical requirements? Nuclear power is so much more efficient than combustion, that it seems to me that if all of your ships have a fission or fusion reactor and your weaponry is based on firing grains of sand at significant fractions of c, your resupply requirements shrinks to fixing things that broke, and feeding the people on the ship (at which point significant portions of your tail are paying for the tail itself!).

    It’s interesting to compare this with the current U.S. Navy: they’ve been investing in sci-fi boondoggles like railguns, directed energy weapons, and in-situ fuel production for decades now, because despite the programs largely being failures (as far as we know…), the savings they would provide would be massive for a fleet that already has lots of nuclear-powered ships with excess power available.

    1. I wouldn’t call DE a boondoggle considering it seems to be showing it’s promise recently, a bit like how ungainly early fire-arms were.
      I would be surprised if EMALS weren’t a direct product of railgun research. The major boondoggles seem to have been poorly thought out littoral ships who were mostly anti-pirate measures that are now secondary to facing off against Chinese cruisers.

      1. The possibility of anti-missile lasers to reliably intercept progressively faster and more maneuverable weapons is just too tempting to pass up and I expect it will continue to soak up gigantic investment until it works.

  15. One thing that wasn’t covered here is how much a sufficient assumption of advanced technology can change the equation, particularly huge increases in the energy density of fuel. I think about something like Star Trek, where they can harness antimatter (about 10^10 times as energy dense as hydrocarbon fuels) and then use that energy for everything from propulsion to food to weapons to entertainment. There are other societal factors at play in Star Trek that would also affect the tooth:tail ratio (eg, despite being the fleet flagship, the Enterprise-D is chock full of biologists and stellar cartographers and children and whatever else), but not having to have transports full of fuel and ammunition following every starship around the galaxy seems like it would change things a lot.

  16. That linked ork logistics guy thread is fantastic. I know you are a historian and so make that your primary argument, but I want to add that it is also entirely appropriate as criticism of the fiction itself. You often argue for the importance of the humanities. Let’s remember that among them literature has value too. And also it’s just plain entertaining to read. Even if I wasn’t learning real history (which I am) nothing wrong with unwinding after a day of work with a fun read about the relative merits of Theoden and Denethor as civil and military leaders.

  17. I have a hard time buying much of the reasoning in this post, because a lot of science-fiction settings seem to assume that air travel is as easy as automobile travel is today, and getting to and from orbit is only slightly harder. Assumptions about FTL (hyperdrives or what have you) are of course highly variable but settings like Star Wars treat even crossing the entire galaxy as relatively trivial.

    On the other hand, the specific comments about Star Wars seem right. While I have zero faith that Lucas actually thought anything through this clearly, you can rationalize a lot of the world-building of Star Wars by assuming the Old Republic was never more than a loose association, that the *reason* it was never more than a loose association is because *nothing else would’ve made any sense*, and Palpatine’s dreams of centralizing authority would’ve been doomed to fail even without any Jedi to oppose him.

    1. Poul Anderson’s Terran Empire stories made that point explicitly for an empire “only” a few hundred light years across. There are a *lot* of stars in that bubble, even discounting the uninhabitable systems, which means some planets’ interaction with the Empire they belong to is the annual ship that shows up to collect tax revenues.

      Though presumably if the taxes aren’t there or are short, the Empire will respond eventually, and likewise if a Merseian raid hits the planet there’ll be a quick appeal for the defense they’re supposedly paying for.

      1. As I recall the numbers (without digging out my books) the Terran Empire was a sphere 400 light years in diameter centered on Sol. There were about 4 million stars in that volume, and the Empire believed that half of them had been visited at least once. Meaning that half of them were completely unexplored.

    2. >I have a hard time buying much of the reasoning
      >in this post, because a lot of science-fiction settings
      >seem to assume that air travel is as easy as auto-
      >mobile travel is today, and getting to and from orbit
      >is only slightly harder. Assumptions about FTL
      >(hyperdrives or what have you) are of course highly
      >variable but settings like Star Wars treat even
      >crossing the entire galaxy as relatively trivial.

      First, not all space opera settings are created equal in this respect. In Star Wars, a teenage farmboy from Tatooine can joke about buying his own used spaceship, and the family has a flying shuttlecraft suitable for bullseye-ing womp rats in. In Star Trek or Dune or Warhammer 40,000, individually operated starships are rarer and harder to find.

      More generally, even if there’s a vast availability of flying, spacefaring, and interstellar vehicles, that doesn’t mean those vehicles do not require maintenance or an industrial support base to sustain them. As Dr. Devereaux pointed out, in real life militaries transitioned from pre-industrial to modern mechanized warfare and the numbers of support personnel climbed tremendously.

      An army that uses trucks to transport its supplies has more “tail” per unit of “tooth” than an army that uses horses, for a variety of reasons… Even though the average citizen of a modern society probably has easier access to automobile travel today than their ancestor of 300 years ago would to horse travel.

      1. “An army that uses trucks to transport its supplies has more “tail” per unit of “tooth” than an army that uses horses,”

        That might be true, but it might not be. It takes a lot more horses than trucks to haul an artillery battery about, the horses take some looking after, and there might not be much grazing for the horses on the Western Front in WW1, or in the Sahara desert in WW2.

      2. The difference is that most of the labor involved in maintaining an army’s horses (mainly grooming and foraging for fodder) was done by the army’s soldiers, acting in a secondary capacity as unskilled laborers. Aristocratic cavalry units might have some civilian grooms to help with that, but the expectation that a cavalryman cared for his own mount(s) is common though not universal throughout history. The same is even more true of artillery horses- the gun crews themselves were doing a lot of the work there.

        By contrast, caring for and maintaining a truck is a specialist job requiring specialized education and tools a frontline soldier cannot carry with him in the field. Thus, an army that relies heavily on trucks will need a corps of mechanics, while an army that relies only on horses will not necessarily need a corps of grooms.

        I observe that your examples of armies that needed to supply vast amounts of fodder for their horses are from the World Wars; this is not a coincidence. These were wars in which the supply chain to the field forces was very much sustained using mechanized equipment. The horses close to the front line in World War One were eating grain shipped to them by railroad and steamship. Horses serving with the armies in North Africa, likewise, with trucks added into the mix. These are not typical representative examples of a muscle-powered army engaged in pre-industrial warfare. As such, they are not very helpful examples to consider when we try to assess what changed between pre- and post-industrial warfare.

  18. “ships in low orbit circle a planet very fast, enabling them to deorbit basically anywhere on very short notice”

    KSP player here. This is only sort-of true. If the destination is on the inclined plane of the orbit (including equatorial), then yeah, you have at most the time it takes to orbit fully to get in position to land on target. If the destination is way off, say on the poles, it’s a long, costly burn to change your orbit to match, and you’d be better served staying at a higher orbit where changes in inclination are cheaper. This is the main reason launch windows to specific orbits are so narrow; the cost to change inclination is extremely high.

    This is all moot if you have exotic tech capable of ignoring inertia or otherwise burning hard for extended periods without turning humans into goo, but if that’s the case there’s no advantage to being in low orbit anyway.

    1. I think this depends what you’re comparing it to. Even a long, slow burn is likely to be faster than the fastest air/ground assault possible. And generally if we’re talking about planetary invasions, then there’s a fleet in orbit, not one ship, so if you’re not in position, it gets assigned to the ship/squadron that is.

      1. “then there’s a fleet in orbit, not one ship, so if you’re not in position, it gets assigned to the ship/squadron that is.”

        The problem there are the narrow windows. For a given target, ships and squadrons are only in position for a few minutes, and then out of position for anything from a hour or so on up to a day or more. The number of ships/squadrons required to cover a even single target can quickly rise to unmanageable numbers as your desired coverage of said target rises.

        1. These are all interesting logistics and physics questions that, sadly, SF authors often ignore.

          What are the re-entry conditions required? If dropships have no problems with re-entry at any speed, the ideal pre-deployment formation is probably to have the fleet spread out at a very high orbit, making maneuvers for a specific landing zone fast and efficient. If re-entry is a problem like it is in reality and trajectory margins are small, you need a much larger number of vessels in a lower orbit and only a small portion will be able to land at a specific location with any speed, and the further off-inclination the target, the longer it will take for the rest of the fleet to maneuver to a safe starting point for de-orbit.

          Ironically, considering these issues makes available many more plot devices with which to pace a planet-side invasion. Spaceflight is hard and unforgiving in reality, and if you need a reason for something to happen in a specific order or at a specific time, there are plenty of real-world factors that could be used to justify it.

    2. Depends on the orbit. If you care about she’s to the ground (either for today’s reconnaissance satellites or for a fleet investing a besieged planet) you’re going to use a polar orbit. With a day that’s substantially longer than one orbital period and minimal cross-range, that gives you landing opportunities twice a day at any spot on the planet, and more often at higher altitudes.

    3. The better way might be to drop into the upper stratosphere and maneuver hypersonically. Maybe change the trajectory then pop back above atmo for a quick suborbital dash before reentering near the intended drop zone.

      Then again, someone with the tech to invade planets might find a few tens of km/s delta v no biggie. At 2g thrust, a dropship can brake from low orbit velocity to literal standstill in about 5 mins.

  19. On Star Wars: It’s worth noting that the Expanded Universe explicitly puts size limits on galactic empires. The Old Republic was unable to make incursions into Hutt Space or Chiss Space because of supply-chain issues; only small isolated patrols could visit there, and only for limited time. Locations of key worlds like Coruscant (administration), Kuat (shipyards), and Kessel (luxury exports) prevented Republican/Imperial control from extending far beyond the Galactic Core.

  20. On Dune: Human society contracted during the Butlerian Jihad. Like any Luddite movement, the Jihad traded automation for labor rights, ensuring that there would always be a need for Mentats. Artillery requires maths; there is some non-trivial trigonometry required to aim over a distance of mere kilometers, and orbital bombardments from above the Kármán line would require dozens of Mentats per operator. While lasguns do not require any such correction and always hit what they point at, they are risky due to the Holtzman effect; lasguns and shields create a mutually-assured-destruction environment.

    The net argument I’m making, I suppose, is that orbital bombardment is a risky, sloppy, expensive method for capturing planets. Without computers, the Holtzman effect is a mysterious barrier which makes military research almost not worth the cost.

    1. If you were firing from orbit there could be more than a thousand kilometres of vacuum between yourself and whatever explosion happened when your lasgun beam reached their shield. The defenders would be a lot closer to their own shield, and therefore the explosion. I would expect the defenders to come off worse.

      Seems perfectly practical to me.

      1. IIRC the lasgun-shield interaction triggers explosions at both the generator and the lasgun, not at the interaction point. It also is noted to have the problem that it may be mistaken for a nuclear detonation with attendant consequences.

          1. I noticed this at no older than fourteen. Put a lasgun on an RPG and nobody will be using shields to protect anything they care about – and certainly not ducal palaces.

            And despite what people say, there’s no huge risk of it being mistaken for nukes more than momentarily and most of the explosions won’t be anywhere near that large (otherwise the trick of leaving shields behind to be ‘found’ by people with lasguns wouldn’t have been viable).

            This is one of the things that makes [i]Dune[/i] unreadable to me these days.

        1. This is one of the parts of Dune that doesn’t really make sense. Why do they even have atomic weapons, when all they have to do to make an atomic-sized explosion is shoot a laser at the enemy? And why are atomics taboo when a cheaper weapon that does the same thing isn’t?

          1. Atomic weapons don’t rely on the enemy having an active shield. If your orbital bombardment plan relies on lasers and the enemy just turn their shields off, nothing happens. A nuclear missile does not depend on the enemy cooperating.

            The problem with laser grenades or missiles in the Dune Imperium is that the explosions, plural because they happen at both shield *and* laser, are effectively nuclear weapons with the yield set by a random number generator.

            If you’re trying to clear insurgents or invaders out of one of your own towns or suburbs, even tactical nuclear weapons are not very helpful. If you’re trying to conquer another industrialised society, blasting everything into rubble with nukes will seriously reduce the property values and GDP.

            And that is with actual nuclear weapons with controllable yields. If your laser missile vs shield might be a grenade, might be a kiloton tactical nuke, might be a megaton city buster, there are very few circumstances where that is a practical weapon.

            The people in the Dune Imperium fight the way they do not just because of technology, but also because of social and cultural factors. They can and do change. By Chapter House of Dune computers are back, so are laser and projectile weapons, and shields are considered obsolete except as lasgun-activated bombs.

      2. Physically practical, yes. Politically / morally, no. The explosion from a lasgun-shield interaction is stated to be very similar to an atomic explosion. The Imperium in the first Dune book operates on the same mutually assured destruction principle as our own, so using lasguns for this kind of orbital bombardment would most likely get your own planet nuked in retaliation.

        (Yes, Duncan Idaho plants a shield to be set off by Hakonnen lasguns in the first book. House Atreides has already lost at that point, he can’t make things any worse. He also (probably) knows by that point about the involvement of the Sardaukar, so the Harkonnens aren’t going to want any attention or investigation of what happens on Arrakis.)

        1. How would retaliation work? The Atreides cannot fire an Interstellar Ballistic Missile back at Giedi Prime – they rely on the Guild for interstellar travel. (And if they could, they could do that if they were facing total defeat anyway.)

          1. The Great Houses have concealed nuclear stockpiles and a vow of collective retaliation; any use of an atomic weapon against humans will result in planetary annihilation to be enforced by all houses. This apparently doesn’t apply if it’s used to demolish a terrain feature.

          2. guy, that strikes me as a very good idea – we could do with something like that on Earth.

            But it is not really needed in a setting where no Great House has a delivery system capable of reaching the home bases of other Great Houses on other planets. This is, however, getting some way from the question of whether orbital bombardment is possible.

            I should point out that you should be able to distinguish between nuclear and lasgun weapons by the decay products of the reaction and telling the Spacing Guild about you lasgun plan during the trip over, and taking a lot of photos of you shooting lasguns at people.

          3. Apparently lasgun-shield interactions also leave behind trace radioactivity that resembles the aftermath of a nuclear strike; the likelyhood of confusion is great enough that it’s explicitly noted to be a deterrent to using remote-controlled lasguns to hit shields.

            As for delivery systems, apparently it is possible to transport atomics via the guild because the threat of annihilation is taken seriously enough no one employs atomics. They’re a party to the treaty banning the use of atomics so they’re probably obligated to assist in its enforcement.

          4. The Atredies do have Interstellar Ballistic Missiles, as do all Great Houses. Indeed getting a hold of such missiles is one of the essential steps that makes you a Great House. These missiles are sited off-planet hidden on asteroids or something. As I recall it, Lady Jessica reflects on this in Children of Dune.

            There is at least a nominal justification for Houses having atomics, that if a hostile alien species turned up someday, the massed Houses could use their atomics to defend against them. (I don’t think there are any known sapient aliens in the Dune universe that don’t ultimately derive from Earth.)

            The Spacing Guild does not have a monopoly on space transport because no one else can build FTL spacecraft, they have a monopoly because only Guild Navigators can make it safe. If 10% of FTL flights are destroyed by collisions with dark nebulas or rogue planets or being eaten by mutant space goats, no one is going to travel that way. Guild Navigators use spice-induced prescience to spot the goats and dodge them.

            But if you are firing planetbuster nuclear missiles, a 10% failure rate is acceptable, you just build extra missiles to be sure one gets through.

            This ties into Tupile. The Harkonnen attack on Arrakis is not how wars are supposed to go in the Corrino Imperium. Inter-house wars are supposed to end when one side has clearly won, and a deal is made that the losing nobles and their families are allowed to go into exile on the sanctuary world of Tupile, so that they are at least alive. In return for this, they don’t press the Big Red Button (BRB).

            But the Harkonnens had a traitor in the Atredies staff who let them get in and grab the Duke before he could press the BRB. So there was no need to let him go into exile on Tupile.

          5. ad (@ad98832376), while it might be technically possible to distinguish between a lasgun/shield explosion and an actual fission/fusion explosion, nobody in the Imperium of Dune has any incentive to let another House get away with it.

            The Great Houses in Dune have a very comfortable lifestyle. Part of that is being able to shield the aristocratic residences and *not* have to worry about you and your family suddenly being vaporised if another House attacks you. The Emperor would no doubt also prefer that when Great Houses fight each other they keep the property damage to a minimum – the Empire wants to collect the same taxes whoever wins.

            So causing atomic scale destruction, even if you don’t actually use an atomic weapon, will result in your evidence being ignored and massive retaliation anyway.

            The Spacing Guild itself is also very comfortable with the current Imperium. So if you tell the Spacing Guild about your lasgun plan, in the best case the ship(s) carrying your force might get “delayed for technical reasons” on the way over while a diplomatic representative from the Emperor’s court arrives to explain to you why your plan is a Really Bad Idea and you should turn around and go home. Worst case, the Spacing Guild might “divert your flight” to say Salusa Secondus where the Sardaukar will make sure you don’t carry out your plan.

          6. “The Atredies do have Interstellar Ballistic Missiles, as do all Great Houses.”

            cptbutton, are you sure about that? As I recall the Atredies and Harkonnenes cannot even put a weather satellite up over Arrakis – but they both have batteries of faster-than-light interstellar missiles?

            I feel it would be a lot simpler to say that orbital bombardments don’t happen in Dune because the Spacing Guild won’t let them. After all, the Guild, like the Shield, was clearly made up by Herbert as a justification for an interstellar empire in which all battles were resolved by swordsman slashing at each other.

            If you start poking at these justifications they start looking a bit fragile because, of course, the justifications were not the thing the author was interested in.

          7. Here is a Google Books image of page 147 of Children Of Dune which covers this. Note I was wrong about it being Lady Jessica in this scene, it is Duncan Idaho.

            “His mind flashed through the review of those conventions which had ended primitive forms of warfare:

            “One-All planets were vulnerable to attack from space; ergo: retaliation/revenge facilities were set up off-planet by by every House Major. Farad’n would know that the Atreides had not omitted this elementary precaution.”

            (If the link works.)


  21. Traveller’s Third Imperium does this pretty well, with a large Empire fielding small mech/power armored forces recruited from a fraction of their inhabited worlds. Mass Effect also implies a low militarization rate for most civilizations.

    I think the kinds of warfare and their objectives also often fall by the wayside when estimating Sci-Fi forces. If you’re dealing in genocidal wars to the last man, woman, or child, your mobilization rates will presumably be very different than if the stakes are a change in elites. Most sci-fi discussion seems to arrive at ludicrous sizes in part because the assumption often seems to be total war. “Why doesn’t the empire build 10 million Star Destroyers” because they’re fighting police actions.

  22. There are a few science fiction writers with a concept of scale, and who make some effort to explain it.
    The works of David Weber would be a good recommendation?

    1. Personally, I think Weber’s books are a good example of why sci-fi writers have no sense of scale. The early Honor Harrington books where she has a small squadron or a single ship had much more engaging battle sequences that the later battles that involve literal hundreds of 8+ megaton capital ships and millions of missiles flying around. Those battles tend to devolve into an impersonal numbers game:

      “The RMN fired X thousand missiles, of which 50% were jammed, 25% were destroyed by countermissiles, 10% were destroyed by PD lasers, and 5% malfunctioned, leaving (X*0.1) thousand missiles to hit and destroy Y superdreadnoughts.”

      Once the shock of the numbers wears off, you’re just reading statistics. And HH has small fleets when compared to his Empire from the Ashes series, from what I gather.

      1. Its that it has hundreds of capital ships and millions of missiles, and that war is often just an impersonal numbers game. From a relatively small base compared to some settings. Manticore is just a single system, but the setting points out that 3 inhabited hi-tech worlds can produce hundreds of capital ships. It even goes into the problems of logistics and having a manpower crisis.

        The Empire series is an even more extreme example, in asking that why can a galactic Empire with thousands of worlds not have nearly a million ships the size of Earth’s Moon?
        It also grasps the concept of time and distance to a far greater extent than most. Earth is initially an unexplored backwater that has a ship on picket duty just because it happens to be on the traditional invasion route of an Achultani “Great Visit”. Due to circumstances, the descendants of its crew find themselves having to fight a genocidal alien invasion with a slither of the resources of the Imperium, while scavenging what they can from its corpse. (Quite tragically and ironically a fraction of this could save Earth, but its all at least a year’s travel away)

    2. Ok, first and foremost, THANK YOU, Dr. Deveraux, for all the great work you’ve done on this blog! First-time commenter, long-time follower here.

      So, without any further ado, I do agree that Weber’s storytelling does suffer a bit when he gets to the climatic battles with a cast in the millions.

      On the other hand, I also think that he does an excellent job of addressing the issue of scale of interstellar war in his Honour Harrington and Starfire books and would definitely recommend them, with caveats. 😉

  23. Dune is my favorite example of the Hilariously Undersized Army. Paul conquers the Imperium, which explicitly spans MULTIPLE GALAXIES, with the Fremen. The Fremen are the only troops worth talking about, given how they curb stomped the Sardaukar, who were the Imperium’s best non-Fremen soldiers. There are about 10 million Fremen: women don’t fight on the front lines, and of course many of the men aren’t of fighting age, so his total strength must be <3 million combatants. Did he have one soldier per inhabited planet. I didn't think the Great Houses were supposed to be THAT pathetic.

    1. Multiple galaxies? Where is that mentioned?

      “Very good, Stil.” Paul glanced at the reels in Korba’s hands. Korba stood with them as though he wished he could drop them and flee. “Statistics: at a conservative estimate, I’ve killed sixty-one billion, sterilized ninety planets, completely demoralized five hundred others. I’ve wiped out the followers of forty religions which had existed since—”

      1. God emperor of Dune, right near the beginning, has Leto musing that he rules over a “Multigalactic empire”. I’m not this level of nerdy over Dune, but I know some people who are, and most of them assume this is an error on Herbert’s part. Pre-scattering, everything in the books that has a location attached to it is in the Orion Arm of the Milky Way galaxy, with the possible exception of Tupile, which is supposed to be really out of the way anyway.

        1. And, of course, we know that Leto’s Imperium can’t have been much bigger than Paul’s, since his whole political program was to prevent migration in order to bottle it up and release it when he died.

    2. As Bret himself has noted, Frank Herbert didn’t try to show exactly how Muad’Dib conquered the Empire.

      My handwave number one: Paul is not destroying the empire, he is taking over as the new emperor. So it isn’t Fremen / Arrakis vs everyone else, it is more like a Roman civil war, or the medieval War of the Roses, with the Fremen vs the aristocratic retainers of the Great Houses. At least some houses could be expected to switch sides, and Paul has enough control over the Spacing Guild to prevent his enemies from combining all their forces against him.

      My handwave number two: the Empire is, early in the first book, described as an unstable political structure by Reverend Mother Mohiam. And at several points in the first book Paul has the sensation that the entire human race is trying to shake things up, that he is struggling to control a vast impersonal force, some sort of teleological evolution. (Not arguing that such a thing exists in real life, just in the book.) So the fall of the Empire is an inevitable social progression / revolution, Paul just happens to be the trigger.

      1. We can also factor in that Paul had the Spacing Guild in pocket with the spice monopoly, and nuclear weapons were freely used during the Jihad.

        So how much conquering and pacification did the Fremen end up having to do, relative to the size of the Empire? When the enemy can freely show up in orbit, without warning, and rain down the Stone Burners and mop up with the infantry.

        Add to this the prescience of their leader, which means Paul knew where his targets were going to be and what they were going to do, and knew who could be cowed and who would pretend to go along only to rise later.

        Complete strategic mobility vs. complete strategic paralysis, and Paul knew what his enemies were planning before they did. He had a strong hand to play, and he played it.

  24. “The Roman legion, for instance, was essentially all tooth”

    In my MA thesis, I showed that Romans seem to have seen more than​ one noncombatant per soldier as excessive, and that Iron Age armies in the western Old World often had between 1 and 4 soldiers per noncombatant

    Which noncombatants get counted and paid out of the central treasury, and which are ‘contractors’ ‘independent businesses’ or ‘personal servants’ is a legal quibble and often reflects some very ugly politics. Poor people curry horses and mend tack so rich people can ride them; women ground flour which men ate; foreign merchants keept the things an army needed coming to it. They often fought (especially in defense of camps, but remember the servants at Bannockburn) and faced as much danger as anyone else: dysentery does not know who pays you. But poor people, women, and foreign merchants don’t put up stelai or write commentarii!

    1. Yeah, there is a point that during the 30-years war there are varying estimates but could sometimes be 4 camp followers for every soldier, if not more. (which to some extent helps to explain the often fantastic estimates fo enemy numbers: Seeing a huge mass of people but not being able to pick out who is a fighter or not)

      I would argue that the difference isn’t neccessarily in tooth-to-tail ratio, but rather that modernity has a tendency to integrate the tail more closely into the military organization.

      1. Xenophon’s second excuse for the Spartan defeat as Leuctra goes as follows (Hellenica 6.4.9) “when both sides were arming themselves and it was already evident that there would be a battle, in the first place, after those who had provided the market and some baggage-carriers and such as did not wish to fight had set out to withdraw from the Boeotian army, (some of the Peloponnesian army) made a circuit and fell upon these people as they were departing, and not only turned them about but chased them back to the camp of the Boeotians. Thereby they made the Boeotian army much larger and more densely massed than it had been before.”

        Modern armies have shops and barbers and cooks (and often live entertainment) but we don’t call them camp followers. Sometimes the cooks get issued rifles and told they need to counterattack, but sometimes camp followers fought too.

  25. Another factor Prof. Devereux doesn’t mention explicitly in the decline of militarization rates is that the “modern system” requires highly-trained soldiers with high physical and mental abilities. An industrial society can’t spare more than a small fraction of its best and brightest for four year stretches. The Russian army, which has demonstrated its complete incapacity for the modern system, may include conscripts serving six month terms; the American army would have no use for such men.

    1. Israel manages putting its best in uniform for 2-4 years just fine; a rich country (unlike Russia) with the political will (unlike Germany or Russia) fighting wars close to its bases (unlike the US) can pull this off just fine.

      In today’s world, that basically means South Korea, frontline NATO states, and a couple of rich and internally strong Middle Eastern states (Israel, Kuwait).

      [Note: I did not know if Kuwait had conscription when I added it to this list. And now I check, of course it does]

      In a sci-fi setting, this would be any rich world that expected to be attacked (ie using conscription to fill out a planetary defense force) or any reasonably wealthy polity on a politically-divided planet (assuming sci-fi technologies make the distances on a planet less logistically significant).

  26. At least for space fleets you have the possibility of “world ships” in which each ship is a more or less self-sufficient industrial civilization, given access to raw materials.

    1. You’re probably going to want to avoid running your primary warships like that; it consumes mass, meaning your ships are either larger or slower than a pure warship of equivalent combat capability. If it’s tactically possible to keep ships nearby but out of the fight you’d want to have your self-sufficent ships have excess capacity to resupply warships, maybe even dock them internally, rather than be warships themselves.

      Of course if your ships are slower than light or your FTL is uncooperative it may not be possible to have resupply ships not vulnerable to attack, at which point making everything self-sufficent makes sense.

      1. Unless you go full Homeworld anyway. Having your fleets that are fully capable of building more flees as long as you find some large rocks or pockets of gas along the way is pretty unstoppable.

        1. Read “Consider Phlebas” by Iain M. Banks. That’s precisely the scenario – every General systems Vehicle contains enough people and machinery and space, to be able to build other GSVs when necessary, and repair any that need to as well. Besides being practically self-sufficient and lugging enough weaponry and people trained in such weaponry, to be a major threat on their ownsome. Though under pressure from the Idirians, they soon develop specialized military vehicles.

          1. I recall that they didn’t even face any real pressure. It was stated to be a war on philosophical and moral grounds, not for practical survival, as the fully self-sufficient spacecraft have no need for contesting “space” or planets. There’s plenty of space and materials everywhere in the universe. If the enemies want them gone they could simply move and not compromise quality of life. Nor do the enemy have anything they want to take. This is even brought up in-universe because of course the Culture is mostly pacifist and some indeed thought it was pointless to fight Idirians over basically no practical reason.

          2. The Idirians liked to conquer other sapient species and force their religion on them. The Culture didn’t like them doing this, and interfered until it led to full scale war.

            If the Culture had been willing to butt out and mind their own business, the Idirians would have ignored them.

            But the Culture’s self-image was of being Benevolent Good Guys who helped and protected the less advanced species. To stand aside and do nothing while the Idirians conquered and converted by the sword would conflict with that self-image. So they fought a war to preserve their sense of their own virtue.

            (Except for the part of the Culture who refused to take part in the war. Being an anarchy, it really couldn’t get consistent behavior.)

        2. Jack Campbell wrote The Lost Fleet. If not capable of full ships, they could do a lot of repairs with sufficient supplies.

          Protecting those capabilities was a major issue.

        3. I’d note that only the mothership, a certain heavily modified mining ship, and carriers have significant onboard production. Most of the fighting is done by pure warships. I think if it’s practical to split production and combat across two ships that’s what would usually be done, and the production ships would hang out a day away from the fighting.

          Major exceptions would be if you can’t keep them out of combat but in support range (your FTL obligates you to travel in groups or has enough of a random factor you can’t guarantee any arrival order) or if you somehow use them during combat, like turning wreckage into fighters in the middle of the fight. Or possibly if various technological constraints are such that one ship is cheaper than two ships half the size.

  27. >why I thin even in a spare-faring context, ground operations might not be dead letter at all.

    Think, I presume?

  28. While unmanned aerial vehicles (“drones”) have been around in warfare for some years now, the current war in Ukraine seems to be showing that they can have a much greater impact that probably a lot of people thought (even consumer-grade drones). If you need an idea for a short post (and feel competent to talk on the subject), I’d be really interested in reading about how drones might change the ways war is conducted in the 21st century and what we might see in the future.

    1. I’ve thought the soldier of the future may be primarily a robot handler: “falcons”, “dogs” and “mules” for automated aerials, attack units and cargo transports respectively.

      1. Software is gonna advance just like hardware, so no need for human handlers either; a deep-learning AI will do better than any human.

      2. Several years ago at work we talked about putting an auto loader on the Abrams. One big drawback was what does the Loader then do? You need four men for security and maintenance (also you need a certain number of people to justify the general’s star) so you can’t get rid of him. At the time drone operator was a possible job.

  29. Most depictions of militaries in science fiction is based on WW 2 tropes that probably won’t hold in the future. Odds are any space faring civilization is going to have drones that are advanced enough they won’t need human troops at all.

    And if they have strong AI those drones aren’t even going to be piloted by people.

      1. In the real world that question is generating a lot of thought and discussion. A book I strongly recommend as an introduction and overview is “Wired for War” by PW Singer, if you want a more academic discussion the book “Armed Drones and the Ethics of War” by C Enemark. (Who as an academic also writes papers.)

        For fiction writers it is simpler. Burnside’s Zeroth Law of space combat: Science fiction fans relate more to human beings than to silicon chips.

      2. Depends on how the strong AI works. Human soldiers might not be vulnerable to countermeasures against AI.

  30. So, what this immediately suggests to me is that Halo’s Grunts should probably never see combat – barring something like a Spartan boarding their ship – because they’re all employed in ‘tail’ making sure that the Elites and Brutes can get into combat.

    Which would probably make the Unggoy very happy.

  31. Your numbers for Soviet population at 1939 are off – at ~168M, it was similar to, but substantially *higher than*, Russia’s population today.

    Perhaps you’re looking at numbers just for the RSFSR?

      1. 1940 would be after the annexation of Eastern Poland, the Baltic states, and Bessarabia.

        I specifically went for 1939 numbers because those territories hadn’t been in the USSR long enough to contribute substantially to the Red Army’s standing force and trained reserves as of mid-1941.

  32. A frightening large number of SF authors live in frighteningly small universes. Discount the galaxy-wide jumps in Isaac Asimov’s Foundation trilogy, and the events take place within walking distance. It should not therefore be surprising that they tend to get the details of military events wrong.

    I’ve played with the idea that a truly ruthless organization approaching the complexity and stupidity of a Galactic Empire, would breed specialized life-forms capable of training to a limited degree, and trained to obey certain classes of people, and predetermined as cannon-fodder. Something between a battledog of the Mediaeval/Renaissance era and the simps of Arthur C. Clarke’s Rendezvous With Rama. Or for that matter, the orcs of Tolkien’s Middle Earth.

    It should be obvious to anyone that as the tail becomes longer, as the tooth’s root becomes longer, that war becomes more total. Once upon a time, villages and cities out of the direct line of march and not closely aligned with the city under attack, would be left alone, because they were largely irrelevant to the slaughter. However, once you have more modern transport and trade systems, once you can interdict from a distance, once you have a more centralized political system where the periphery is as much a threat as the current target, war becomes closer to total, and citizens are considered fair prey.

    That’s something that Stanislaw Lem considered worthy of several novels and short stories, Fiasco being perhaps the most obvious of them all.

    1. I would have thought that the smaller the scale of politics, the more total the average war becomes, simply because that way a larger fraction of people are close to the frontier and likely to be closely involved in the war. The further you are from the war zone, the less of a concern it is.

    2. Isaac Asimov said in one of his essays that his ideal would be that everywhere in the world was in walking distance of a teleportation booth. A lot of his fiction reflects this.

  33. In the recommended article on Pydna, I assume when they mention camp sizes of 650 m^2 or 700 m^2, they actually mean (650 m)^2 or (700 m)^2.

  34. Another factor that might push a planetary defense force vs elite assault army split is the simple question of transport. It depends on the setting, obviously, but in most lifting say 5% of a planetary population of a major world (what I’ve seen quoted as a high end for mobilization) at once is simply not happening, there aren’t enough ships. You’d thus want a very well-trained force with high-quality weaponry, though probably skewing physically light for what you have the capacity to manufacture, for offensive operations. Meanwhile defensively you’ve got an entire planetary population to draw upon and your main size/mass constraints are the square-cube law and hiding from orbital bombardment.

    That puts a lot of pressure on the attacker to get the most out of their orbital control and attempt to use mobility to defeat enemy forces in detail, which in turn means the defenders are going to want to establish a network of strongpoints each durable enough to see off the entire enemy force hurling out of orbit; they’ll have serious trouble reinforcing each other because large movements of troops will be exposed to bombardment and the attackers will pick orbits that let them exploit that. The strongpoints will have to be covered against the same either by technological defenses, buried shelters, or just being co-located with something the attacker wants intact-ish with horrific consequences if the attacker decides they don’t want it intact that badly.

    1. See Arthur C. Clarke’s “Earthlight” for the sort of ship vs. ground duel this can entail.

      1. Was that the one with the battle on the moon where the moon installation “slid” as the ground it was sitting on boiled into lava?

  35. Have you considered moving to Substack?

    You do not need to switch immediately. Just copy your articles over to that platform and see how many paying subscribers you will get. My understanding is that the academic job market has not been kind to you; and it is absolutely worth trying out whether Substack can improve your income situation.

  36. “I will note that while Cameron describes it as ‘astonishing” to find that most captives were women and children this is actually a very typical pattern in raiding both because adult males are the most able to escape a raid but even moreso because the adult males are killed rather than taken captive.”

    both here and in the “that dothraki horde” series you mention that this was the typical pattern for raids by the AmerIndian tribes. yet in the dothrki series you called the assumption by white settlers that the Indians were going to rape the women racist. despite saying that the goal of raiding was to steal the women to marry into the tribe to grow the tribe. these marriages wouldn’t have been willing on the women’s part. we have a word for that, “rape”

    how can you say that was a racist assumption by the white settlers when that was literally what the plains indian tribes were doing? forced marriage is in fact, rape

    1. I don’t think this is an accurate description of the comments he made? In part IV, there’s a brief discussion of rape in American Western films, but no comparison is made.

      He makes two statements in III which are sort of in the ballpark of what you’re referring to:

      “Consequently, compared to the other cultures of Westeros and Essos, the Dothraki are disproportionately associated with rape and sexual violence; it is hard not to notice how this plays into bad old cringe inducing Hollywood stereotypes of Native Americans as a threat to white women on the frontier.”

      “The idea, advanced by Martin, that the truly stunning amount of rape – most of it not in the context of war – in A Song of Ice and Fire somehow reflects medieval social norms or a true vision of the past or particular cultures is to be rejected. Needless to say, it is not considered OK to just begin raping dancing women in either Great Plains Native American or Mongol cultures.”

      I’m not seeing anywhere he takes the view that he was saying that kidnapping and ‘marrying’ (raping) captives was not rape or not present? Do you have a quote or citation to him making such a claim?

      One problem with the stereotype isn’t that raiders weren’t a threat to women, it’s that it overlooks the fact that the vast majority of the violence over the period was flowing American->Native American and casting that as ‘reprisals for raids’ is ahistorical and pretty silly. It also tends to flatten the Native Americans into ‘raiders’ (as I accidently did in the first sentence of this paragraph).

      1. This may seem a strange question, but can it be easily demonstrated that the vast majority of the violence over the period was flowing American->Native American? The settlers clearly had more permanent victories than the Native Americans, but that is not the same thing as killing more people. After all, the Germans killed a lot more Soviet citizens than vice versa in 1941-1945, but that doesn’t mean they made more permanent conquests.

        1. So, this isn’t my area of expertise, but my understanding is that it’s very difficult to get good casualty figures, in part due to the complicating factor of disease and propaganda (going both ways, ie ‘the evil Indians butchered X hundred civilians’ and ‘we defeated X thousand enemy warriors’). However, casualty figures aren’t actually the controlling factor, at least for what I meant by the direction of violence.

          The scenario very much wasn’t what you sometimes see on frontiers: raid->counterattack, refortify frontier. It was a constant expansion into new territory for explicitly expansionistic reasons. If you show up, armed and ready to fight and claim new territory, then when the locals raid you to try to drive you out, you have instigated that violence, even if they shoot (as it were) first and regardless of the total number of casualties on either side.

          1. A lot of the effects of violence are indirect – you kill adults then the old and the children people starve; you drive off herds, or destroy gardens, or destroy shelter or food stocks then people starve or die of exposure. You make life so unsafe they move, then they die on the move. Of the 27 million Soviet dead, 9 million were military casualties (and 3 million of those were POWs deliberately starved or left to die in exposed camps); most of the other 18 million were not actually shot (or, if Jewish, gassed) – they died of the many ways war kills. Same for frontier wars.

          2. There is indeed a sense in which the settlers were responsible for all violence involving settlers by definition; but there is also a sense in which if someone squatted on my land, and I responded by killing him and kidnapping and raping his wife, the violence was flowing from me to the squatters.

          3. Various indigenous tribes were frequently at war or raiding each other. In terms of moving into new land and driving off the prior inhabitants, the settlers weren’t doing anything that the prior inhabitants weren’t already doing to each other, the settlers just did it better and backed by the US army.

          4. However, a large chunk of the tribal violence was the tribes just keeping their old traditions of raiding their neighbors regardless of race. You can only claim that it was intruding on their land if it was their land. (And a far amount of land was sold to whites, or unoccupied because of disease or raiding by other tribes.)

            The Seminoles got into quite a snit that after they moved to Florida and violently claimed land, the Spaniards claimed title on the basis on having bought it from the tribes they violently dispossessed, but by any rational standards, the Spaniards were in the right.

          5. A lot of responses here boil down to ‘raids are bad for the raided,’ which is true and ‘native Americans engaged in warfare too’ which is also true.

            But neither is relevant to the question of ‘who is instigating a conflict between Europeans and native Americans’ The answer has generally been the Europeans. You don’t conquer continents by accident. Now, there were certainly exceptions and there were plenty of intra-Native American conflicts, but that’s not really particularly relevant?

            My point is not that Native Americans were precious, peaceful people until Europeans ruined everything. It’s that stereotyping entire cultures as vicious raiders, while ignoring the fact that the people who are being treated as innocent victims are only present there because they are part of an invading and colonizing effort, is an extremely limited viewpoint.

            And that’s without getting into the history of direct violence and sexual violence directed towards Native Americans. I mean, we want to talk about killing children there’s plenty of history of that on all sides, along with kidnapping, rape and indoctrination/forced assimilation. The problem with the stereotype isn’t ‘raiders were really peaceful and no threat’ but that:

            1) Native Americans were not solely raiders
            2) Warfare was very definitely not going only one way (or at least very definitely not going only ‘innocent colonists abused by vicious raiders’)
            3) If you invading/colonizing an area, you are, in fact, starting it, as it were. And though that doesn’t justify war crimes, or other immoral acts, there’s a long and ugly history of invading a place and then claiming that people you’re invading deserve it because of how they’re resisting your invasion. We can see this in the present day, in some areas and we’ve certainly seen it historically.

          6. You lump together everything that occurred as a single conflict? Not reasonable. Not unless you are going to argue for ethnic cleansing.

          7. “You don’t conquer continents by accident.” Actually we sort of did. Repeatedly during both the British colonial era and the USA era the central governments kept trying to set borders via treaty to “solve” the problem of perennial conflict with various indigenous tribes. Only to have land-hungry settlers ignore what the guvumint pen-pushers said and illegally push the frontier. Even if it meant fighting literally scores of often bloody Indian wars, with the central government more or less dragged into the conflict by their obligation to protect the settlers. To some extent the frontiersmen and settlers were just another “tribe” in conflict with the indigenous ones, albeit a tribe with a huge number of allies to the east.

          8. It is very possible, and indeed sensible, to tally up a large number of similar and closely associated trees and start thinking of them as a single “forest.”

            There can be plenty of instructive information to be had from examining the trees. One should not lose sight of the individual trees. But it would be unwise to say “there is no forest here, only a lot of individual trees that just coincidentally happen to carpet the land for miles around in every direction.”

            As noted, one does not conquer a continental landmass in its entirety by accident. It is downright disingenuous to say “there was no conquest, just a long succession of isolated individual fights where we kept marching from coast to coast for 250 years, endlessly defeating a bunch of vicious raiders who totally had it coming every 50-100 miles along the way.”

            What is actually relevant to the discussion at hand is to point out that in Dr. Devereaux’s criticism of Martin’s portrayal of the Dothraki, the specific point where Dothraki-as-rapists is condemned is one thing, not another thing.

            It is not that Dr. Devereaux is arguing “it is counterfactual and also racist to portray pastoral nomads as committing rape during their raid-centered style of warfare, or as coercing enslaved women into sex after capturing them in warfare.”

            It is that Dr. Devereaux is arguing “it is counterfactual and also racist when pastoral nomads are portrayed as doing almost nothing but rape and warfare, to the point where rapes seem to be the main form of sexuality and warfare seems to be the main form of human interaction among them.”

            The Dothraki aren’t just committing rape (also murder, torture, arson, etc.) during raids and warfare. They are committing rape (and murder, less so the other things) all the time. Which is a counterfactual (and also racist) exaggeration of the real world behavior of groups such as the Mongols, Turks, and Plains Indians.

            To be sure, the idea that a Dothraki wedding is “a dull affair” unless at least three murders take place can reasonably be called exaggeration or a comment on the scale of elite warlords‘ weddings. But it illustrates the scope of the problem. Which is, again, not that the Dothraki are portrayed as doing vicious and ugly things during war as such (though the destructiveness and purposelessness of those wars are exaggerated). It is that the Dothraki are portrayed as a culture so singlemindedly focused on viciousness and ugliness that we see no other facet of their culture. And that they are so ‘warlike’ that they seem to have almost no capacity for normal things human societies engage in among themselves, such as “arts and crafts” or “consensual sex” or “resolving disputes without murdering each other.”

          9. The tribes in the Americas when Columbus arrived were not the first inhabitants. Those survived only on the southernmost tip of South America.

            Since that can’t have been by accident by your argument, they must have planned genocide.

          10. It’s fair to say that you don’t conquer a continent entirely by accident, yes, but the idea that there was some kind of actual plan beyond “Hey, wouldn’t it be great if we held everything between the Atlantic and Pacific oceans, we should do something about that” just doesn’t hold up.

          11. Pick a cold planet without a biosphere. Since you’re going to be terraforming that much.

          12. Whoops. The last one is a confusion on my part.

            I meant to observe that in actuality, some people thought that and often weren’t in a position to implement it.

          13. Were genocides committed before Europeans arrived? Almost certainly? Have you read some of the stuff our host has referenced on the ‘cutting off’ way of war?


            Very ugly shit.

            Which, again, not relevant?

            Anymore than ‘hey, look what the U.S. did, clearly it’s fine if Britain reconquers us?’

            As for a plan…yes, there very definitely was? The United States repeatedly purchased/acquired by treaty vast tracts of land from other European powers, then set about conquering the actual inhabitants. There wasn’t some ‘these are the exact borders we’re going for, we’ll set up our lines of advance like this’ plan. But the expansion of the United States was in fact an extremely planned/supported endeavor.

            We didn’t ‘accidentally’ do the Louisiana Purchase or the other acquisitions (see e.g Expansionism and manifest destiny were major (though not uncontested) strands of political theory and practice in the early US.

        2. There’s also that violence from natives is documented – violence against natives much less so. Natives raping settlers is a horrendous crime, settlers raping natives is a sport (same as blacks and whites in the South). The difference is not in methods but in attitudes – the settlers were (for the most part) avowedly genocidal – they wanted the land, but not the inhabitants.

          1. And left no documentation of those activities? Of which they were not at all ashamed. I find that off.

          2. @Roxana: In fairness, there is usually a certain amount of “what happens on campaign stays on campaign” that goes on with the details of this stuff, and particularly egregious crimes were usually hushed up–for example, the Gnadenhutten massacre, in which a group of Pennsylvania militia heroically slaughtered a bunch of Indians who’d converted to Christianity and foresworn violence, was only talked about by the Moravian missionaries whose converts were killed.

          3. “And left no documentation of those activities?”

            Have you looked for such documentation? Some does indeed, exist, see for example Michele de Cuneo’s “Letter to a Friend”.

            More generally, any conflict that ends say, “Other Pequots were enslaved and shipped to Bermuda or the West Indies, or were forced to become household slaves in English households in Connecticut and Massachusetts Bay” ( is saying ‘and at least the women amongst them were almost certainly raped.’

            Again, not my area, but some superficial googling of reasonable sources very much does not support the notion that there’s no documentation of this sort of thing.

    2. It’s not just Amerindians. All raiders be they Achaeans hitting Mediterranean cities, Vikings hitting Northern European monasteries and settlements. Steppe horsemen or Africans or New Guineans, women are captured, enslaved and usually raped at some point. Women are loot. That’s a cross cultural fact of life.

      1. The Greeks and Romans, being as binormative as always, seemed to have that view on young male captives too. Sallust has Caesar mention “the rape of maidens and boys” among the horrors of war, and Xenophon writes in the Anabasis of a boy-loving soldier who made him ask the Thracian king Seuthes to save a handsome young enemy combatant

  37. I want to push back somewhat against the comments about militarization rates. They’re falling, yes, but that’s because there has not been WW2-scale total war. Smaller countries involved in acute total wars have had some high militarization rates. To wit, Israel militarized around 10% of its population in 1967 and 1973, and that’s with around 20% of the population (Arabs and Haredis) not serving and to a large extent not integrated with the national civilian economy.

    Today, at a time of peace, the IDF is around 2% of Israeli population, or around 2.6% omitting Arabs and Haredis (Arab and Haredi civilian workforce participation is higher now but still far below mainline Jewish levels). South Korea’s ongoing militarization rate is 1%. The limiting factor is that these countries’ populations expect a mostly civilian economy, but in a symmetric war the rates can spike as the reserves are recalled; Finland, with shorter conscription tours of duty than Israel and Korea, has around 16% of its population in the reserves.

    I think the US also overcounts its tail. The CBO guideline counts anything outside a BCT as support; CABs and special forces count as combat troops but the support ratio is for BCTs only. If you only count combat companies in BCTs (artillery, cavalry, armor, infantry, and UAVs, but no HQs or support, or let’s also say not engineers even though they do forward-deploy), it’s 1,790 for an ABCT, 2,076 for an SBCT, and 2,027 for an IBCT. So you get 60,810 US Army-wide tooth troops in BCTs alone, and more like 67,000 with engineers, and then you should add choppers (around 30,000 tooth + tail) and special forces (another 30,000). The ratio looks more like WW2’s 1:4 than like 1:8.

    Then there’s the issue of scale. Larger armies benefit from scale better – they only need one general staff no matter the size. It’s a big discussion topic in Europe, where one of the points made by proponents of an EU army is that there’s 27-plication of command and headquarters functions. The EU’s deployable strength is lower than the US’s even with a nominal numerical advantage, because having 27 armies means there’s no scale, and for example each army has a separate equipment pool with its dedicated maintenance staff.

    1. I suspect the existence of nuclear weapons has something to do with the modern low militarisation rates. To quote Sir Humphrey Appleby:

      “Conventional forces are terrible expensive Prime Minister. Much cheaper just to press a button.”

      I suppose the EU problem is that you need to have a single foreign policy, if you are going to have a single Army to serve it. I get the impression that the foreign policies of France, Germany and the East European member states have not been perfectly aligned over a matter of existential importance to the Eastern members.

    2. There is also the fact that the US has converted many previously military positions to civilians or contractors. Base security, public works, janitorial, motor pool – most of these are done by civilians or contractors at established bases.

      1. On the other hand, there’s a reason for that under the modern system of warfare.

        Soldiers will of course be expected to do a certain amount of maintenance work on their own equipment, to maintain personal hygiene, and so on. But in a large conscript army where every draftee is theoretically part of the “tooth” if they are assigned to a unit anywhere near the front lines… Well, there are a lot of draftees sitting around, and you might as well get some use out of their labor. The soldiers who are nominally part of the “tooth” spend some proportion of their time doing ‘kitchen patrol’ and cleaning up their own barracks, same as the soldiers in the “tail” of the army.

        One book I often recommend as giving certain insights to the physicality of early Industrial and Early Modern warfare is John D. Billings’ Hard Tack and Coffee, recounting the author’s experiences in the Army of the Potomac outside of combat. Actual battles are rarely discussed, and the focus of the narrative is mostly on more mundane considerations such as how soldiers ate, slept, sheltered, maintained their clothing, and so on. One concludes that the men of the Army of the Potomac did a lot of manual labor- because to a large extent they were responsible for cooking their own food, maintaining their own clothes, and erecting their own shelter, especially when out in the field on campaign.

        On the other hand, the Army of the Potomac had a much simpler training regimen than is typical of a modern army. Even the kind of X-week boot camp found in modern armies would probably be overtraining by their standards, and the sort of continuous, ongoing training with various equipment and practices that modern armies do would be almost unheard of. Most of the inexperience and lack of ‘training’ commented on in Hard Tack and Coffee is, ironically, inexperience with civilian activities! Soldiers would come prepared for war, in some cases even overprepared and loaded down with extra weapons gifted to them by relatives… Only to prove inexperienced at maintaining their own clothes with needle and thread, or ignorant of how to properly care for horses.

        So somewhere in the transition between the US Army of 1862 and the US Army of 2022, it became necessary for soldiers to spend a lot more of their time practicing with fiddly bits of equipment and sitting in classrooms studying field manuals. Correspondingly, this meant a lot less of those same soldiers’ time was available to do things like cook their food, maintain their clothes, or build and clean out the buildings they worked in. Which, in turn, meant outsourcing more of those tasks to either noncombatants, as the people who were trained as combatants could no longer do it themselves and still be fully competent as combatants.

        1. Actually one of the main reasons those sots were converted was money. Congress wanted those jobs to help their districts. Their is talk of moving some back to government jobs, such as housing oversight, as the contractor are not doing a good job

          1. Partly as @scifihughf says, and partly something else…

            First, the process I describe at least partly predates the political process you describe. My point was about how the armed forces have embraced far more extensive division of labor as a corollary of the twentieth century’s changes to warfare.

            In modern warfare, the main thing that determines victory (outside COIN warfare) is the ability to bring firepower to bear. Things that permit the delivery of firepower at the price of requiring huge support staff are usually effective, and the use of weapons usually requires extensive training.

            Thus, there is every incentive to shrink and ‘professionalize’ the actual combat arms, while establishing divisions of labor. Whether the people who labor to support the combat arms are civilian contractors or recruits within the military who spend all their time is a bureaucratic detail. Either way, you see more “tail” per unit of “tooth,” if you’re just counting people who fight directly. The most obvious example is in aerial forces where you may need dozens of people to support a flying vehicle that only actually enables one or two people to fight.

            By contrast, in the Early Modern system of warfare, lower levels of firepower mean that you need to pack a lot of armed men into your army who actually fight. Barring outliers with a huge technological disparity, sheer number of soldiers is generally more important than the quality of armament you bring to bear. There is, in this system, no way for ten men with a complex weapon backed by 100 technicians to defeat 110 men armed with simple weapons they can carry and maintain personally.

            Furthermore, the basic sustenance requirements of an army are still difficult to provide. Even when (as an Early Modern general) you are routinely accepting low standards of housing and food quality by modern standards, it’s still a major logistics challenge just to keep troops fed and watered even in temperate climates.

            These two factors combine to incentivize the highest possible tooth-to-tail ratio, insofar as “tooth” is defined as “people who, when battle comes, will definitely be fighting.” In this system, it is generally desirable to take all but the most specialized jobs and have them be performed directly by the soldiers of the army. The more soldiers you have making themselves useful doing unskilled manual labor to assist with the army’s sustenance requirements (foraging, maintaining shelter and clothing, and so on), the easier it is to sustain a large group of fighting men.

            The US military originated under the latter system of warfare, and transitioned into the former.

        2. The conversion of various support roles to civilian contractors does not change the practical tooth:tail ratio for the US Army, although it does obscure things a bit. The motivation for contractors may have been money and jobs for congressional districts, but those tail requirements already existed, they weren’t being created just for contractors.

          As Simon_Jester explains, the US (and other modern) military found that ‘tooth’ soldiers no longer had enough time to do most of their own ‘tail’ support. So in the first half of the 20th C, and still in many militaries today, there are a lot of people who have military rank, are subject to military discipline, and get paid military salaries but have ‘tail’ roles not very different to civilian.

          You can give the appearance of shrinking your military, and generate income for the private sector, by converting those roles to contractors. Whether this actually improves the effectiveness of your military is open to question: see this report about what happened to the US backed Afghan military when their support contractors decided to seek other opportunities:

  38. “On the flip side the defender is likely to have at least numerical parity if simply because they don’t need to fly through space to get to the battlefield.”

    I think that this is implicitly assuming that interplanetary transport costs are high, or the war between a handful of planets. But if costs are low, and the contending space opera empires large, a single planet might be attacked by an army of billions.

    The defender would also face a very unfortunate geometry. It would not be like defending an island, where you can withdraw into the interior of the island. People do not live in the interior of a planet. Everything they want to defend is on the surface of the planet. The defence usually wants to be between the thing to be defended, and the enemy. How does a ground-bound army interpose itself between people/ cities / etc on the ground, and an attacker in the sky?

    And unless the defender has some interstellar equivalent to an ICBM, a disappointed attacker would always have the option of twirling his moustache, crying “If I can’t have it no one shall!” and calling for an exterminatus. He might be unhappy about that conclusion, but probably not as unhappy as the defender, and it might have a helpful effect on other planets attacked later.

  39. “One of my favorite bits of math”

    Alas, that only works if you assume the oceans are colonized with floating townhouses or something. 500e6 km2 * 400 people/km2 = 200 billion. If you stick with land area, 150e6 km2, it’s more like 70 billion people. Also, while the Netherlands exports net food monetary value, I don’t know if it’s self-sufficient in calories; Wiki rates it at only 54%, though a LEI economist claims it could be sufficient if it needed to be. France is sufficient, but only 100+ people/km2; Germany is close, 80%, and 233 people/km2, so that could give 35 billion people. (Assuming the Earth was terraformed to be as habitable as Germany.)

    Bangladesh claims to be self-sufficient, and has a density of over 1000 people/km2; that would give 150 billion people. If you colonize the oceans too that’s 500 billion people.

    Oceans are important, I think the bulk of our oxygen comes from plankton there… OTOH oceans are kind of nutrient deserts, so in theory could be better managed.

    1. (Assuming the Earth was terraformed to be as habitable as Germany.)

      Do you need terraforming make the average patch of terrestrial land as habitable as Germany? I would think you can not get more than a few hundred miles further from the equator than Germany and still have productive agriculture, but you can get thousands of miles closer to the equator than Germany and still have productive agriculture

      1. Yes, but there are very large fractions of the Earth’s landmass that, while closer to the equator than Germany, are also a desert or a mountain range- unsuitable for agriculture. Or are a tropical jungle with thin soils- unsuitable for agriculture without extensive work to cut down the jungle without utterly destroying and depleting the soil.

        The question then becomes, what percentage of Germany is arable land, compared to what percentage of the world as a whole? Would extensive alterations to the Earth’s terrain be needed to make the whole Earth, proportionately speaking, as arable as Germany?

        1. I think it is worth remembering that most human being live in South or East Asia, where you can grow a lot of rice. And that has been true for thousands of years. That is the most habitable part of the world. Europe, the Americas and Africa are historically marginal land.

    2. Of course, Bangladesh is not exactly famous for having highly reliable food security capable of sustaining its population, though that reputation may be undeserved in the modern day. Also, Bangladesh is, as I recall, mostly low-lying river valley, the kind of terrain famous for exceptional fertility.

      The big question is whether densely populated areas (like Bangladesh or the Netherlands) are ecologically self-sufficient. That is, not only is the population physically growing enough calories, but could they continue to do so without, for example, unsustainably importing fertilizer from somewhere else? A single country the size of the Netherlands can get away with feeding itself using phosphate fertilizers strip-mined out of a desert in Morocco. But if the entire world were populated to the Netherlands’ level of density, there wouldn’t be enough phosphate rock lying around to last us all very long.

      Ecologically sustainable carrying capacity of a planet is of course a complicated question. And to some extent it depends on the question of “how willing are you to turn the planet’s ecosystem into an artificial life support framework for its human society?”

      1. If you are building a world city, you are plainly willing to overwrite the world’s natural ecosystem. Additionally, if you are an SF empire building a world city, you are probably capable of recycling your phosphorous and other materials used in agriculture, and may well be growing via hydroponics and other methods that don’t really require soil.

        If you’re building multi kilometer high cities over an entire planet, I’d say the local ecology is screwed, and the only reason that you can’t stack 100 levels of hydrophonic gardens at the bottom of every structure is that the waste heat from the sun-lamps would cook everyone.

          1. You may need some biosphere for the oxygen atmosphere. Even with all those gardens, it could still take millions of years to generate one from scratch (might be less since you don’t have oceans).

            But, if you are designing a galactic capital planet (call it Washington DC), then putting it on a frozen ball would make sense, but most of the time the assumption is that the world city “just grew” after it was made the capital for other reasons. In which case cold and uninhabitable to start with is not an option.

            There’s always Niven’s Puppeteers solution (also used in World out of TIme by the same author IIRC), simply move the capital planet further out from the sun. This is almost certainly doable for a galactic empire.

  40. I’ve generally envisioned spave empires being very feudal. The logistics of controlling a planet that doesn’t want to be controlled make it almost impossible. (The defender will always have more material since their supplies won’t first have to escape their own planet and then go through whatever version of FTL your universe has) . So wars are small fights between small elite space navies over the miniscule amount of goods that it makes sense to transport between worlds.

    1. I think there’s a pretty big range of space travel difficulty where interstellar warfare in the form of sending an army from one planet to another is basically flatly impossible unless there’s a massive population or tech imbalance, like thousands to one or tank-to-bronze-age level. In such a case I’d expect interstellar empires to rely heavily on local popular support if they want to maintain control beyond “send tribute up the elevator and we will not drop asteroids on you”. Which may very well be a satisfactory level of control to lots of space warlords.

      If they want to exert more control than can be accomplished by the threat of dropping rocks*, I’d think they’d rely very heavily on information war to persuade the local population and hopefully decent chunks of the military to launch an uprising, then supplement it with seizing orbital control and sending down small units of elite special forces and the most mass-efficent tanks and aerospace craft money can buy to support their local allies. Plus potentially targeted dropping of rocks on military installations if it won’t antagonize their local supporters, because as long as they hold the orbit there’s no risk of mutually assured destruction.

      *For which the main limiting factors are that if you tell people to do something but can’t verify whether they’ve actually done it they can just not, and as your demands get more extortionate they may at some point decide to call your bluff, and then you must either let them get away with it or commence actually dropping rocks.

      1. This also reminds me of one thing I’ve long thought about interstellar trade, especially if you’re stuck slowboating it and especially if there are aliens: the most convienent interstellar trade good is intellectual property. Technical designs or art, either has value hugely out of proportion to its mass and if you’re contacting an alien civilization you’ll have totally disjoint cultural backgrounds and can swap microfilm of the Lovure for advanced technical data or a few space iPhones. Even if it’s not to their psychological tastes it still has value as a curiosity like deep learning art does to us and one copy can be shared by their whole planet, so a hold full of art is worth much more than its weight in gold or even quality electronics.

  41. I can’t disagree with your conclusion that USSR of 1940 had a sorta similar total population to the Russian Federation of today, but I thing the actual numbers were roughly 195 million then vs. 140 million now. Moreover, the USSR’s population in 1940 was much younger than Russia’s is today. In the 18-24 age group, Russia’s pop. is now about the same as the UK’s, 1/3 that of the US and 1/15 China.

    1. Agreed. This statement ” especially when you keep in mind that the USSR of 1940 had a similar, but lower total population to the Russian Federation today (c. 110m to 140m)” is definitely wrong, by a long way. In fact the Russian Empire even in 1914 had a bigger population than modern Russia, and its military-age population was far larger.

  42. I find the idea of inter-planetary combat fascinating, and the aspect that most Sci-Fi seems to ignore is how impossible it is to hide in space. In 2017, the VLBA telescope made out the Voyager transmitter ( – basically equivalent to seeing a flashlight from 11 billion km away (interplanetary distances tend to be more like 100 million km, so that is a long distance even by interplanetary scale). Some sort of cloaking would likely help, but you are still dealing with the fact that space is extremely empty so it would be difficult/impossible to become completely invisible, even at long range.

    To make matters worse, this visibility would apply to enemy fires (missiles, rockets, bullets) as well. To avoid elimination, space warships would have to be equipped with extremely capable movement systems or extensive counter-measures. I’m no military expert, but I wonder to what extent pitch-battles would even occur in space, compared to the possibility of firing smart missiles from the other side of the solar system.

    The only upside is that angular resolution is quite limited. The Hubble Telescope is an above-average telescope (to put it mildly) and it only has a resolution of 0.04 arc-seconds, which can only resolve objects larger than 10-100 km at typical inter-planetary distances. In other words, all space vessels will look like dots and the only differentiating factor will be how it moves, how bright it is, and what spectrum of light it emits/reflects. The most obvious solution I can think of to the long-distance fires problem is deploying a lot of “dummy” vessels to conceal which vessels are your active warships.

    Either way, I would be interested in a book that takes a more hard-science approach to space combat. Most books I have read tend to mimic terrestrial combat (some combination of naval or air warfare) or they employ “shields” to force the vessels in close. But it seems to me that space combat would have it’s own unique set of challenges that are quite different from naval/air warfare.

    1. If you’re firing from four light-hours away, and the other side can see your missiles as well as you can see them, their movement systems don’t have to be all that capable and they will have a very long time to prepare to intercept them. I’m very doubtful there will be much meaningful ship-to-ship combat at ranges longer than a light-minute unless the stealth problem is solved for missiles somehow. There’s no atmosphere or fog to distort anti-missile lasers and a generous window to launch interceptors.

      If anything, I expect combat will mostly occur at “close” ranges of a couple hundred thousand kilometers, ranging up to a few light-seconds if ship-to-ship lasers dominate. Dumbfire kinetics and lasers aren’t going to be terribly effective if you’re firing from far enough away the ship has spent several seconds randomizing its acceleration before the light from it reaches you. You could do smart missiles, but they’d have to be armored to resist lasers on final approach, have defenses of their own to protect against interceptor missiles that have all the same tracking features and require much less fuel, carry sensors advanced enough to track the ships that might launch flares and powered chaff and evade kinetic interceptors using the missile’s own momentum against it, and at that point the distinction between a “missile” and a “drone ship with a warhead” gets pretty fuzzy. Depending on various factors you might have them strike directly, or you might mount railguns to fire from effective railgun range (when you hit if the ship maxes/zeroes its acceleration on detecting the railgun firing) or lasers from effective laser range (similar; they don’t see the shot but they can plug an RNG into their engine controls at ten light seconds) or fire sub-missiles as it reaches effective laser/railgun range so it takes multiple shots to destroy all of them, or even have a single missile loaded dead center to fire off at a thousand kilometers so you can eventually pick the main body up if fuel is enough cheaper than ships.

      1. You make a good point about lasers and the usefulness of random acceleration against dumbfires, I hadn’t considered that. However, I’m skeptical of the value of mounting railguns/lasers versus a warhead or even a simple kinetic weight. Obviously, if your goal is to have a ship in orbit around a target at the end you can’t kamikaze it, so guns/lasers/etc… will still have a place. But kinetic impacts would be even more effective in space, as you have more space to accelerate to extremely high velocities. To make matters worse, I would hypothesize that a ship coming in from outside the planetary orbit would have maneuver advantage, as only a small delta-V from the incoming ship will radically change where they impact/pass by the planet, while a ship in low orbit around a planet requires massive delta-V to get the same orbit change (from Kerbal space program, for major orbit changes it is more delta-v efficient to reach escape velocity and return than to attempt the orbit modification directly).

        1. Well, depending on how good the lasers are it’s entirely possible they could melt an incoming missile’s engines a minute of flight time away and render evasion completely trivial, and if your long-range missile drives are relatively comparable to ship drives and railguns give much higher acceleration it’s possible that you could reliably get a missile within twenty thousand kilometers of your target but not reliably into your target but a well-timed railgun shot could connect.

          Mounting lasers gives you a weapon that crosses light-seconds in seconds and not hours to days (A light-second is a really long distance), so you can fire and plausibly hit a randomly evading target from much further than you could with a kinetic projectile, and where a missile would spend hours getting shot with a laser, so if your lasers do ship-killing amounts of damage they’re an extremely potent weapon. Kinetic weapons might do much more damage, but in ship-to-ship combat you just care about doing enough damage to render the target nonfunctional and you can use kinetics for hitting the planetary defense centers with their three meters of reinforced concrete layered atop them.

    2. I’m reminded of David Weber’s series of Honor Harrington novels (and the tie-in novels by Eric Flint). Over the course of twenty-five years of near-continuous warfare between the Manticoran Star Kingdom and the Republic of Haven, several revolutionary technological advances are made. Shielding goes from having several exploitable holes to nearly complete coverage. An ally’s so-primitive-no-one-realized-what-it-could-do engine technology made carriers with assault craft superior to dreadnaughts. And a simple FTL jump detector was eventually refined into a high-bandwidth data system operating at Cx64. The latter especially makes missiles receiving FTL-updated targeting and evasion data all but omnipotent against STL systems. As a result, when Manticore and Haven make peace and take up a common cause against the huge but corrupt Solarian League, the Solarians discover to their dismay that they’re fielding the equivalent of a 1915 military against a 1945 military.

    3. “The only upside is that angular resolution is quite limited. The Hubble Telescope is an above-average telescope (to put it mildly) and it only has a resolution of 0.04 arc-seconds”

      Whether Hubble is above-average, average, or below average depends on the metric chosen for comparison… And in the case of angular resolution, Hubble is hobbled by it’s (relatively) small mirror diameter and limited light gathering area. (Angular resolution is proportional to wavelength/diameter.)

      The 200 inch Palomar Telescope could theoretically manage .02 arc seconds in visible light, but it’s actual performance is limited to around 30 arc seconds by a combination of it’s 1940’s construction and atmospheric distortion. One can easily imagine an orbital long baseline interferometer array beating that (theoretical) resolution by a wide margin.

    4. Deep space (interplanetary) battles have a lot of problems if you’re using plausible rockets. If pre-modern bluewater naval battles were unlikely because you couldn’t *find* the enemy, deep space running battles will be hard because it’s hard to *meet* the enemy. Because the closest analog for spaceships isn’t boats or aircraft, it’s *trains*, despite the lack of physical track.

      Why do I say that? Going fast with high-thrust rockets takes a lot of reaction mass. Civilian ships would probably launch from A, go to B, and expect to refuel there. Just adding a round-trip capability gets you into exponential territory. Free-form high-speed maneuvering is basically impossible.

      Stuff like ion or plasma drives do potentially have high delta-vee and room for a lot of course changes, but they’re low thrust/acceleration, so such changes will be slow — probably too slow for military purposes.

      So imagine an attack fleet going from Mars to Earth, and Earth wants to meet it away from Earth. To have a running battle, the Earth fleet needs to accelerate, decelerate to a stop, then accelerate the other way in such a way that the Mars fleet just catches up to them and you can duke it out. Or do a fast flyby of the Mars fleet like jousting, then turn around, but that’s even worse because now you have to catch up to the Mars fleet from behind, then slow down again so you don’t pass them. And you can’t keep zipping past them back and forth because you won’t have the delta-vee to do so.

      (And I might be ignoring the whole role of gravity and orbital mechanics, too.)

      Just in terms of reaction mass, you could send 4x as many missiles, that either hit or fly off into the avoid, as running-battle ships.

      As rocketpunk-manifesto pointed out, you can have something like running battles, but they won’t be in deep space, they’ll be in orbit. Which is probably near the thing you’re fighting over anyway. Orbit means you get repeated passes for free, it also means that you have a big thing occluding your view, giving room for, if not stealth per se, then information gaps.

      The planet isn’t defenseless either; lasers and particle beams don’t care about the gravity well, and do care about heat sinks, which the planet has way more of. Submarines might move around underwater stealthily, then pop up to shoot lasers before submerging again, shielded from tracking and even nuclear retaliation by thick layers of water. And “dropping rocks” isn’t that easy or quick.

      But rocks might give a justification for a deep space running battle: say you’re genocidal and want to move an asteroid onto a planet. Well, that’s pretty visible, so the defender fleet would come and adjust the trajectory again. So your fleet is still there to defend the asteroid. So you get a big fight over control of the asteroid, far from the planet, and the asteroid itself isn’t big enough for useful gravity.

      But if you want to conquer the planet, then defense is probably shooting missiles at you and waiting for you to try to enter orbit.

      1. I think any combat that occurs in between planets will be exclusively of the high-speed pass variety unless either both sides think matching velocities is a good idea or one side has run out of reaction mass and the other has plenty and wants to match. That said, it’s possible the window for engagement on the pass will be hours long and ships will fire off their entire stock of ammo during it.

        Any such engagements would presumably occur because the defender does not want the attacker to reach orbit even briefly, and get within optimal (minimum interception window) range for “light” planetary bombardment. A 15-kiloton nuke aimed at the civilian fusion power plants that have been cut over from the main grid to powering your ship-annihilating death laser will ruin a lot of people’s days.

      2. This is also a good point, and it makes me wonder if planetary (or lunar/asteroid) installations might end up playing a major rule. They would be easier to fortify, by simply digging underground, and make it easier to conceal your firepower and store the massive amounts of fuel required to launch missiles and ships.

  43. It’s interesting. As our host says, the logistical constraints of interplanetary invasion would be likely to impose unheard-of tooth-tail ratios on interstellar militaries: 1:100, 1:1000, etc., do not seem implausible.

    A historical comparison occurs: the “tooth-tail” ratio of Apollo 11 was 1:200,000—400,000 people working over the course of a decade to arrange an unopposed landing of a “half-strength infantry fireteam” on the single most convenient astrophysical target available, for a mission lasting about a hundred and thirty five minutes.

    Of course, our hypothetical space empire can probably do better—Apollo 12 would have displayed dramatically better ratios already—but perhaps not better enough to claw back all those orders of magnitude.

    But I think space opera seldom takes this into account because it (by necessity) upends the narrative expectations of war stories. The tooth-soldier sitting at the far end of a 1:1000 ratio is just a very different person than the military characters we are used to seeing. Or perhaps rather a very specialized subset, in the way that astronauts were sub-selected from within military aviation.

    No grunts at the pointy end of a 1:1000 tooth/tail ratio. No stormtroopers who can’t shoot straight—wash them back out into the tail! Wash them out entirely! One guy losing his footing and Wilhelm-screaming down a bottomless pit represents the effective loss of a Roman legion. The scene in Star Wars where our heroes swing across the gap would be a greater cost to the Empire than the Varian Disaster to Rome. (Not proportionally, but still.)

    So, no goons. No average joes. No fresh recruits. No human waves. Something very strange, and focused, and a little unearthly. (Pardon the pun.)

    If we go through the common examples listed above, the ratios seem to suggest a soldier more Sardaukar than Stormtrooper—elite, obsessed, fanatical. More Space Marine than Sardaukar—heavily, heavily equipped, a postcyborg military aristocrat who expects to deploy with a handful of peers and take a mode of zero casualties per mission.

    But perhaps more Jedi Knight than any of these. The Force is, of course, fantastical—but if it existed, if Force sensitivity were one in a million, if you took children and trained them from birth to use a power the rest of the universe could not resist—that would give you the sort of person it would make sense to spend ~15,000 work hours (of highly skilled labor!) putting on-target for a quarter of an afternoon. (Work hour stat from here:

    It’s hard to conceive of what this would look like in a less-fantastical context. Powered armor is probably only just the start. Drone swarm command? Network attacks? Combined arms with orbital and de-orbiting fires? etc.

    Interesting to think about. And interesting to imagine what stories might emerge. Or which could be told in a tail of such improbable size.

    1. AIUI, a lot of the high cost of space missions is because they tend to be one-off (or two-off) prototypes, so you’re paying for design each time with little re-use. So just doing the same thing over and over could knock an order or two off.

      1. The high cost of space flight is due to a wide variety of things… Most of which are due to spaceflight’s utter immaturity compared to other forms of transport.

        But yeah, overall, assuming that spaceflight is going to remain hideously expensive and difficult (it’s already a fraction of Apollo’s) is betting against history and current trends. That’s not a good way to bet.

        1. Oh, sure, I’m not betting on the Apollo ratio being sustained. But it also seems plausible to me that the floor is quite a bit higher than the current floor (which is itself rising)—hence the proposal of scenarios in the 1:100 or 1:1000 range. Can sci-fi societies do better? Depends on the technology to hand, of course. If they can space-jaunt, e.g., there’s no reason interplanetary invasion would be any harder than any other kind. But I think that the high numbers are at least conceivable, and they would create the opportunity to tell different stories than the ones we tell today.

      2. IF it works (we’ll see), SpaceX’s Starship could drop the price of to-orbit to somewhere under $100 a kilogram, with potentially even lower costs if enough economy of scale happens.

        1. Starship’s system could work absolutely perfectly – and SpaceX could still go bankrupt. Because to drop costs to orbit, it has to fly frequently *and* fully laden.

          And it’s not at all clear that sufficient demand exists in the near to medium term. Which means someone has to eat the operational costs until the demand does materialize. (Which could take a decade or more.)

          1. SpaceX plans to use Starship to boost Starlink satellites. Those are intended to be periodically replaced and they’re conducive to being launched in large batches that can make use of Starship’s capacity. Plus, if it costs thousands rather than millions of dollars to launch a payload, satellites don’t have to be multi-million dollar extravaganzas anymore, and more cheaper ones can be launched.

          2. At best, Starlink stems the bleeding and modestly slows down the rate at which bankruptcy approaches. It’ll only take 30 launches to replace the *entire* planned constellation…. Even if the birds have an on-orbit life of a year (which is ridiculously low), that’s less than one launch a week. That’s simply not enough.

            That’s the basic problem with Starship’s economics – it’s designed around airliner like demand and flight rates. But airliner like demand doesn’t currently exist.

            In theory, the cost of satellites will drop dramatically… In practice? That remains to be seen. Satellites are expensive for a wide variety of reasons, and launch costs (and weight restrictions) are only one of them.

          3. Currently the only cheap satellites are microsats that take advantage of being charged only the marginal cost of being added to a launch that will take place anyway. Primary payloads that cost less than the cost of getting them into orbit aren’t worth it; reduction of launch costs open up new vistas.

          4. I have no familiarity with Starship whatsoever, but it’s going to have to lower launch costs before people start making tons of cheap satellites. If it needs to be frequently launched fully laden to get costs to $100/kg, then its costs aren’t going to be $100/kg and it won’t make sense to create satellites that are only worth it at that price.

            If the technology works, though, it would be a valid buisness strategy to run it at a loss for a decade so that people start making lots of cheap satellites and it becomes profitable at that price. That of course assumes you have the money to sustain that at least until people start building and are willing to reserve launch slots.

          5. Remember that Elon Musk’s ultimate goal is to make spaceflight cheap and high-capacity enough that settling Mars becomes feasible. If the means to do so can at least partially pay for itself then fine, but he wants to do it anyway. For instance, I don’t know if the decades and billions spent developing the Falcon 9 has turned a net profit or not, but it has made spaceflight much cheaper and more routine, and served as an invaluable test bed for the technology needed to make Starship possible. Although technologies ultimately have to pay for themselves, some degree of “because it’s there” and “if you build it they will come” is often needed to kick-start the process.

        2. I fully expect technological improvements and economies of scale will bring down prices, but $100/kg is still a lot when you’re trying to lift an armored division and I think there’d be a serious effort undertaken to get the best combat effectiveness to mass ratio possible.

Leave a Reply