The first thing to note is that when we talk of “Roman roads”, we almost always mean the viae publicae, roads built by public officials (Initially censors who let out the contracts to build such public works, although later roadways get named after the consuls and praetors who constructed them as the Romans build more of them) and was maintained by the state. But of course these major state highways existed within a wider network of local roads (a via vinciales or actus, pl. actus – everybody loves the Fourth Declension! It’s a step in the right direction) which might or might not be private (a privatum iter). That distinction is important, because it wasn’t that all Roman roads were of the high quality we tend to think of – the roads we’re thinking of were prestige projects undertaken by the state, but a whole [mess] of private lanes and dirt paths existed too.
That said, for the major viae publicae, the combination of archaeologically preserved examples and a few references in literary sources gives us a good sense of what “best practices” for Roman road construction were. To start with, obstructions were cleared and then a trench was dug where the road would go. The trench was then filled with three packed layers: a stone layer, then a gravel layer and finally a sand or cement layer. On this was placed the surface of stone blocks, cambered so that rainwater drained to the sides of the road. The stone surface was then held in place by an umbo (literally a “knob” or “swelling”, the same word is used for shield bosses) on the edges of the road. Within a town, the umbo might in turn border a stone sidewalk, but out in the countryside, that wasn’t present.
On the one hand, we know that the high-quality construction of Roman paved roads impressed in the ancient world; Dionysius of Halicarnassus proclaims them one of Rome’s three most magnificent works (alongside aqueducts and sewers, Dion. Hal. Ant. Rom. 3.67.5; note also in this vein Strabo 5.3.8; Plin. HN 36.125). On the other hand, we also know that not all Roman roads were built to the same high standard. While we know that various Roman legislation spelled out required road widths, for instance, the actual width of Roman roads varied widely. And while we’re here, I should note, the story you have heard that modern roadway width or standard railway gauge being based on Roman road width (or, heaven help us, Roman “war chariots”, a thing the Romans did not have) is rather wrong. Nor were they always well-constructed or maintained; Julian, for instance, bemoans a poorly maintained road in a letter to Libanius in 363.
Also, if I may stop a moment, one thing I hear frequently said about Roman roads is that their survival serves as some sort of indictment on modern road construction, “they don’t build them like they used to!” There are a few layered bad assumptions here. The first, of course, is survivorship bias; that road outside of Chalcis that Julian complained was sinking into a swamp probably isn’t available for us to see and similarly badly maintained or poorly constructed Roman roads are simply gone or only visible with archaeological methods and so unavailable for comparison. At the same time, modern roads are asked to do things which ancient roads were not; I do not suspect many Roman roads would last very long if hundreds of 20-ton (or more) trucks were rumbling over it daily. Indeed, preserved Roman roads outside of the cities often have deep ruts worn several inches deep in the stone from the passage of carts (carts and other vehicles were generally banned inside of Roman cities; city streets were for pedestrians). Finally we know that Roman roads, just like modern ones, required maintenance and reconstruction fairly regularly in the period of their use; Laurence (op. cit., 66) has a neat table of inscribed milestones on the Via Appia, for instance, referring to repair or reconstruction, with sections of the roadway repaired in stages from 97 to 110AD.
Nevertheless, the generally high quality of Roman roads, as noted, was recognized in antiquity and is still apparent in their survival today. In contrast to many unpaved sections of Persian royal roads, for instance, Roman viae publicae were paved as a matter of course and standard widths (c. 3.5m wide; the oldest Roman laws, the Twelve Tables, set a minimum width on straights of c. 2.4m though as noted this was not always followed) were fairly generous. And they tended to be well-engineered, with relatively flat surfaces (which are often quite a bit less flat today due to ground shifting as well as the erosion of concrete between paving stones) and straight lines, though the notion that all Roman roads were ruler-straight is, of course, wrong (though some are!). In particular, the Roman system of road construction, while it demanded considerable up-front labor costs, was seemingly designed to keep long-term maintenance demands low. This was a style of road building which accepted big up-front costs in exchange for lower long-term maintenance, which of course demands a lot of initial state capacity to manage the costs and labor demands (something Rome’s successors would mostly lack).
So as roads go, the Roman ones were uncommonly good (albeit with a wider degree of variation than is often appreciated), but the marvel of the Roman roads is not that they had a few good, paved roads – states had been building paved roads for some time, as noted above – but that they had a massive system of them. Scale, more than quality, was the Roman achievement here, though the quality was also quite high.
Bret Devereaux, “Collections: Roman Roads”, A Collection of Unmitigated Pedantry, 2023-06-02.
August 28, 2026
QotD: Roman roads
August 22, 2026
QotD: Fortified towns in the Middle Ages
Nevertheless, the larger size of the area being defended makes it harder to deploy concentric defense-in-depth solutions. There are also transit considerations: towns are, after all, economic centers, typically sitting at the junction points of many different local and regional routes, which means they tend to need lots of entry-points to facilitate all of that travel and trade and in turn that means lots of potentially vulnerable gatehouses. Meanwhile, internally, the desire for residents to move around the town freely and general land-use considerations make subdividing towns into walled districts (very common in video-games for memory-management and loading reasons) very difficult and as a result very rare. You can see this, for instance, looking back at the period city-maps from the Lonely Cities series: a good number of cities have either a citadel or a section of the town which is walled but also outside of the main circuit (often separated by a river), but full concentric styles of defense or multiple cell-like walled districts don’t really occur. There are exceptions, but they are relatively rare: most cities have a single primary circuit of walls, which, if they are breached generally give access to the main area of the town.
In Europe, the most complete and formidable set of city walls were without question the Theodosian Walls built to defend Constantinople during the reign of Theodosius II (r. 402-450). The Theodosian Walls stood for the whole Middle Ages but they really reflect the culmination of the ancient system of city wall fortifications, the ultimate answer to the kind of “big army” playbook the Assyrians (and Romans) employed. So I want to both talk about the defense system but also note points where its method diverges from the walls of later medieval towns.
The Theodosian Walls had to cover some 5.7km and had nine gates; both numbers should immediately make clear that you needed quite a few people to man these defenses effectively. This was in particular a factor for town defense: most of the defenders are going to be citizens of the town called up to defend it in an emergency, rather than a standing guard. As far as I can tell, this seems to have made towns somewhat more vulnerable to surprise “hasty assaults”, if the attacker could force a gate along a long circuit before the defenders knew where the attack was coming. Of course that problem is intensified by the fact that this is a single-point-of-failure system: the loss of one gate against a superior foe generally meant the loss of the town.
The Theodosian Walls are unusual in that they are a set of triple walls. The main fighting position was on the inner wall; in front of this was the peribolos, a low, flat, unobstructed gap, after which was the lower outer wall, a secondary forward fighting platform. The towers of the inner wall project out into the peribolos for exactly the same reason castle towers project: defenders on those towers can fire, unobstructed, down the peribolos, meaning that an attacker can find no safety here. The crenellation of the outer wall likewise only faced out, so attackers who seized the outer wall would have no cover from defenders fighting from the inner wall. Next, after another gap was a simple low forward wall, little more than a stone breastwork about 1.5m high. Finally, the low wall stood at the inner edge of a large moat, some 20m wide and 10m deep, because no good defensive system is complete without a ditch.
That low wall is a kind of what we call an “outwork”. This particular kind of outwork is a product of the ancient world and uncommon in medieval fortifications to my knowledge; it was a response to the increasing use of torsion catapults during the Hellenistic period. While it did offer a position for defenders to use to sally out (for instance to destroy enemy siege engines), the primary purpose was to prevent enemy catapults from being able to direct fire at the base of the main wall and its towers. Even if knocked down, such an outwork’s rubble would obstruct direct fire on the curtain wall or the towers. This idea – using an outwork to physically block lines of fire on the curtain wall – is going to come back a bit later.
In contrast, most later medieval town wall circuits generally only have a single curtain wall, supported by a moat. As with castles, the wall is supported by projecting towers which can direct fire along its length and hardened gatehouses. Where cities sometimes do employ defense-in-depth it is with a “citadel” – a hardened fallback position either inside the city or (more often) built in to the wall circuit; essentially a complete castle built into the city wall. The role of the citadel is important for thinking about what these defenses are intended to achieve: defenders who have lost the main area of the city but have held out in the citadel are unlikely to be in a position to recapture the city or even to seriously threaten to do so. But the citadel has important negotiating power. Even a small number of surviving defenders can likely hold a citadel (with its stockpile of supplies) for a long time; taking such a citadel by storm poses all of the problems of storming a castle, with the added disadvantage of having to do so through city streets. Consequently, defenders holed up in the citadel are in a good position to negotiate for terms: perhaps being able to march out unharmed in exchange for turning over the citadel, or else negotiate less onerous terms for the city in general (alternately, they might try to hold out until a friendly relief army could arrive).
And that brings the focus to something important about almost all medieval siegecraft: the siege and negotiations generally proceeded side-by-side. By raising the cost of a siege (either a blockade or a storming assault), the defender could improve their bargaining position, even if they were always going to have to surrender. Meanwhile the attacker, by very visibly preparing for a deliberate assault could signal to the defenders their resolve and thus attempt to encourage surrender. As far as I can tell, far more fortified medieval towns and castles surrendered than were ever taken by storm and so consequently a lot of this fortification work is as much about shifting bargaining positions in the event of an attack as it is resisting an attack.
Bret Devereaux, “Collections: Fortification, Part III: Castling”, A Collection of Unmitigated Pedantry, 2021-12-10.
August 16, 2026
QotD: The “Marian reforms” were not a thing
- Cohorts: Experimented with before Marius, especially in Spain. Marius uses cohorts, but there’s no evidence he systematized or standardized this or was particularly new or unusual in doing so. Probably the actual breakpoint here is the Social War.
- Poor Volunteers Instead of Conscripted Assidui: Marius does not represent a break in the normal function of the Roman dilectus but a continuation of the Roman tradition of taking volunteers or dipping into the capite censi in a crisis. The traditional Roman conscription system functions for decades after Marius and a full professional army doesn’t emerge until Augustus.
- Discharge bonuses or land as a regular feature of Roman service: Once again, this isn’t Marius but Imperator Caesar Augustus who does this. Rewarding soldiers with loot and using conquered lands to form colonies wasn’t new and Marius doesn’t standardize it, Augustus does.
- No More equites and velites: No reason in the source to suppose Marius does this and plenty of reasons to suppose he doesn’t. Both velites and equites seem to continue at least a little bit into the first century. Fully replacing these roles with auxilia is once again a job for our man, Imperator Caesar Augustus, divi filius, pater patriae, reformer of armies, gestae of res, and all the rest.
- State-Supplied Equipment: No evidence in the sources. This shift is happening but is not associated with Marius. In any event, the conformity of imperial pay records with Polybius’ system of deductions for the second century BC suggests no major, clean break in the system.
- A New Sort of Pilum: No evidence, probably didn’t exist, made up by Plutarch or his sources. Roman pilum design is shifting, but not in the ways Plutarch suggests. If a Marian pilum did exist, the idea didn’t stick.
- Aquila Standards: Eagle standards pre-date Marius and non-eagle standards post-date him, but this may be one thing he actually does do, amplifying the importance of the eagle as the primary standard of the legion.
- The sarcina and furca and making Roman soldiers carry things: By no means new to Marius. This is a topos of Roman commanders before and after Marius. There is no reason to suppose he was unusual in this regard.
So the Marian reforms … were not a thing. Functionally none of what is described as happening in them was new or unique to Marius. Indeed, the most substantial reforms are either things that were already changing (and which Marius seems to have had little role in) or things which had not yet changed but which would, under Augustus. Indeed, one of the problems with the assumption of a Marian reform is that it takes a whole lot of changes which were more likely a package of reforms under Augustus and pulls them forward in time to Marius, doing some damage to our understanding of both figures.
But Augustus makes much more sense as the figure doing many of those organizational changes. For one we have sources actually telling us he did them, from the standardization of military service (Dio 54.25.6), the creation of retirement bonuses and the aerarium militare to fund them (Dio 54.25.6 again, but also Res Gestae 17), and the radical expansion and formalization of the auxilia as part of the Roman army (Tac. Ann. 4.5). The one thing missing is citizenship-for-service, which we can see emerging under Tiberius, the next emperor immediately after Augustus. And it makes sense because Augustus is, by necessity, doing a bunch of other things with the legions too. He’s instituting a whole new command structure, with dedicated legionary commanders (the legatus legionis) serving under his own provincial commanders (legati Augusti) with imperial procuratores handling the former role of the quaestors. Also, the legions get their citizen cavalry (though a quite small detachment of it) back too.
So instead of thinking, “Marian reforms” – which were, I must stress, not a thing – you ought to be thinking about a period of tactical, organizational and institutional change beginning in the second century, accelerating in the first century and then finally being instituted as a comprehensive set of reforms and formalization by Augustus, not by Marius, which codified a lot of change that had already happened over that long period. A long process with a punctuation mark at the end rather than a singular moment of reform associated with a singular Roman general.
Alright. Now one of you go and fix the Wikipedia entry.
Bret Devereaux, “Collections: The Marian Reforms Weren’t a Thing”, A Collection of Unmitigated Pedantry, 2023-06-30.
August 10, 2026
QotD: Cleopatra’s relationship with Egypt and her Egyptian subjects
The factors usually pointed at to demonstrate Cleopatra’s attachment to Egypt are that she learned the language, engaged with Egyptian religion and represented herself in Egyptian fashion in royal artwork and commemoration meant for consumption in Egypt. And that’s true, she did those things! But the context often missing from using that as a clear indicator of Cleopatra’s cultural “Egyptianness” is that, apart from learning the language, those were common things for Ptolemaic rulers to do, despite the fact that the dynasty maintained a pronounced ethnic hierarchy in Egypt and aren’t generally regarded as being particularly attached to its people or culture. Nevertheless, there’s not nothing here in the sense that Cleopatra showed – or at least wanted to show – more than the normal Ptolemaic attachment to Egypt as a place and the Egyptians as people, which is to say almost any at all.
Let’s start with languages, because I think this fact can be presented in a somewhat distorted way. The language of the Ptolemaic court was Greek, initially Macedonian Greek (the Macedonians had a pronounced accent), though Plutarch notes that some of the later Ptolemies had lost their Macedonian accent (Plut. Ant. 27.3-4). Cleopatra, by contrast, was the first of the Ptolemies to bother to learn Egyptian (which should tell you something about the character of Ptolemaic rule; imagine if King Charles was the first English king since George I and kings from the House of Hanover to bother to learn English). The problem with this fact is that it is incomplete, presenting Cleopatra as a Greek-speaker who learned the language of her people out of sincere devotion, but that’s not what Plutarch says. Plutarch says:
She could turn [her voice] easily to whichever language she wished and she conversed with few barbarians entirely through an interpreter, and she gave her decisions herself to most of them, including Ethiopians, Troglodytes, Hebrews, Arabians, Syrians, Medes and Parthians. She is said to have learned the languages of many others also, although the kings before her did not undertake to learn the Egyptian language, even though some of them had abandoned the Macedonian dialect.1
So let’s unpack that. This isn’t a native speaker of Greek who learned just the language of her subjects, but a spectacularly skilled linguist who learned a lot of different languages, quite regardless of if she ruled the people in question. Running through the list, she evidently learned Ethiopian, the language of the people on her southern border, the speech of the Troglodytae, the people who lived on the coast of the Red Sea (a hinterland of her kingdom). The “language of the Hebrews” here is probably Aramaic rather than Hebrew (which would also cover much of Syria), while the language of the Medes and Parthians might mean both Old Persian and the Parthian language. To which we must add Egyptian, implied by that last sentence; it also seems fairly clear Cleopatra knew at least some Latin.2 This is part of why I find arguments that use Cleopatra’s knowledge of Egyptian as strong proof either for her Egyptian ancestry or deep attachment to Egypt less than fully compelling; she was surely not Parthian and did not have a deep attachment to Parthia, but she learned their language too. Again, there’s not nothing here, but it’s not a slam dunk either.
What of Cleopatra’s role in Egyptian religion? The recent Netflix miniseries makes quite a fuss about how Cleopatra’s first action as queen – that we know of, because this is simply the first moment she is visible in the sources to us at all – is to partake in a religious festival installing a new sacred bull in a shrine at Hermonthis, close to Thebes in Upper Egypt and how this was indicative of a deep love of Egyptian religion. And it may well have been, but it’s also worth noting that Ptolemaic rulers had been taking part in Egyptian religion for a long time. Most of the massive temple complex at Philae, for instance – even farther from Alexandria than Hermonthis, I might add – was built in the Ptolemaic period, with major expansions by Ptolemy II and III. Even more relevantly to our subject, this was a [devotion] to Isis; Cleopatra was not the first Ptolemy by any means to claim that goddess for the dynasty.3
Nor had this even been a new practice for the Ptolemies. Alexander was presumably crowned pharaoh in Memphis in the traditional manner, visited the oracle of Amun at Siwa and dedicated a major set of construction works in the temple of Amun at Thebes. The Ptolemies followed this model. Ptolemy II went to Mendes to personally conduct the ritual of the process of the Ram of Mendes and then extended the temple, he presided over the inauguration of a temple at Pithom, and so on.4 And on it went with Ptolemy III and on down the line (indeed, becoming a bit more pronounced as we move into the later Ptolemies, beginning with Ptolemy V Epiphanes, potentially as a result of the increasing weakness of the Ptolemaic dynasty both internally and abroad). Ptolemaic kings were regularly in and out of Memphis, the religious capital of Egypt, building new Egyptian temples, attending Egyptian religious rituals and so on. The reason is that the job of Pharaoh was as much (if not more) a religious job as it was a political one and it was a job that only the Pharaoh could conduct. And the idea that a Ptolemaic pharaoh with a legitimacy problem might consecrate some temples to help fix that wasn’t new either; Ptolemy VIII Physcon is out doing that thing in the mid-second century. Consequently, involvement in Egyptian religion was a requirement of the position, quite apart from the question of how much the holder of that position cared.
Cleopatra, like any Ptolemaic ruler, engaged in this sort of building, though it seems striking how much of her building program was composed of Hellenistic structures in Alexandria, even as Duanne Roller admits, “the queen was not a major builder”.5 She restored the gymnasium of Alexandria, made major repairs to the Lighthouse, built a new precinct in honor of Julius Caesar called the Kaisareion, and apparently built a monumental tomb for herself, also in Alexandria. Examples of her building outside of Alexandria are fairly thin: the completion of a temple at Dendera begun by Ptolemy XII, extensions to the temple at Hermonthis (where she visited early in her reign), never entirely finished and a temple to Isis in Ptolemais Hermiou in Upper Egypt. Cleopatra could hardly be accounted the great renovator of Egyptian religion or religious sites.
Likewise the assertion that Cleopatra represented herself in Egyptian artwork in Egyptian style and dress; that too was true of all of the Ptolemaic rulers from the very beginning. Indeed, Cleopatra’s self-representation as Isis was very much not new, but in fact the normal way Ptolemaic queens were represented (to match Ptolemaic kings being represented as Osiris). Here, for instance, is Ptolemy II Philadelphos (r. 284-246) with his queen Arsinoe II represented in Egyptian style on a relief now in the British Museum.
Like all Ptolemaic rulers before her, Cleopatra presented herself in an Egyptian form, conforming to Egyptian style, in Egyptian contexts. And, like all of the Ptolemaic rulers before her, she presented herself in a Hellenistic form in Alexandria and on her coinage. While you will see people point to various busts and frescos as being images of Cleopatra, it is only her coins where we can say for certain that an effort was made to capture her likeness (her Egyptian portraiture, though labeled, follows Egyptian stylistics, not her actual likeness); we have no idea if, for instance, later Roman artists had any information about what she actually looked like or if the busts we suppose to be of her actually are. Be wary also of Greek and Roman frescos showing fair-skinned women reported to be Cleopatra; almost all women in Greek and Roman fresco are shown as fair-skinned; it’s a standard artistic convention. Coins, of course, record no skin color and neither do our sources for Cleopatra; anyone who thinks they can tell you what color her skin was is unduly certain.6
Now this isn’t to say Cleopatra is doing nothing new. She does seem to have been somewhat more involved in Egyptian religion than her predecessors, though the exact degree to which this is true is hard to see because we don’t have anything like a full accounting of their actions; if it is a difference it is a different in quantity, not kind. Certainly her building program wasn’t anywhere close to the scale of the early Ptolemies, but perhaps she took more direct personal involvement. Once again, the unknowns here are formidable. Dorothy J. Thompson in the the Cambridge Ancient History7 takes this, along with Cleopatra switching her regnal name from “philopator” (“loves-father”) to “philopatris” (“loves-homeland”8) to be indicative of a clear effort by Cleopatra to be, “indeed queen of Egypt”, particularly in the context of dynastic struggles with her brothers: outflanking them by getting the favor of the people.
The problem with this narrative is the timing: Cleopatra makes that switch in titles in 36/5 BC, which is to say in year 16 of her 21 year reign, eight or nine years after her last serious competition for the throne, her youngest brother, had died. Instead, Bingen9 argues that the homeland Cleopatra is professing love for is the greater Macedonian empire – the empire of Alexander – which she may have seemed in 36 on the verge of reclaiming.10 Antonius was just about to give her (through her children) sweeping portions of Roman Syria, Cilicia, Cyrenaica and Armenia; he would also give them titles to Medea and Parthia, which he didn’t control (though perhaps in 36 he hoped to) but would represent collectively the whole of the old Seleucid Empire near its height. Bingen notes the other change to Cleopatra’s titulature at this point was rather than just being “the goddess who loves her father” she becomes “the young goddess who loves her father and her country”; the added νεωτέρα (“young”) perhaps being intended as a link to Cleopatra Thea, Seleucid queen-consort of Syria from 150 to 126 BC and Cleopatra’s own great-grand-aunt. Cleopatra, after all, was about to achieve the long Ptolemaic dream of controlling all of Syria.
Strikingly the sources do not tell us that Cleopatra was seen as any more distinctively Egyptian than her father or that she had some groundswell of popular support in Egypt. Indeed, Cassius Dio implies the opposite (Cass. Dio 42.34-6) that the Egyptians favored Ptolemy XIII in the succession dispute and it was only the force of Caesar’s arms that turned the issue to Cleopatra. Of course we have to be somewhat skeptical of this account too: Dio is marinating in a literary tradition that is very hostile to Cleopatra and also makes no effort to distinguish between the Alexandrians and the rest of Egypt, two very distinct groups. In any case it is clear that Cleopatra achieved no groundswell of support in Egypt in those early years: she had to raise her army in Syria and in the event it wasn’t the stronger force as she and Caesar spent the winter of 48/7 besieged in Alexandria by Ptolemy XIII’s then larger army (Caesar had not brought his full field army), until the arrival of an army out of Roman Syria led by a collection of Caesar’s allies tipped the balance. Julius Caesar was never a particularly cautious commander, so if he looked at whatever force Cleopatra still had and said, “we need to wait for reinforcements”, it could not, at that point, have been very substantial. Afterwards, Caesar would leave four legions in Egypt to secure Cleopatra’s rule, which seems not to indicate her great popularity or any great faith in her ability to hold the kingdom. After 44, she’s able to rule without direct support, but of course after 44, she and her children are the only dynastic options left, her brothers and sisters all being dead (some at her hand). Finally, when Octavian comes for her, there is no major effort by anyone to stop him militarily. If Cleopatra had the love of the Egyptian people at that point, we sure don’t see it.
In my view then, it seems like an argument can be made that Cleopatra had a fairly modest propaganda effort towards presenting herself as more in touch with Egyptian religion, but the extent and quality was at most only a modest improvement over the previous Ptolemies and apart from keeping her on the throne (with Roman backing), it doesn’t seem to have achieved much.
Bret Devereaux, “Collections: On the Reign of Cleopatra”, A Collection of Unmitigated Pedantry, 2023-05-26.
- Trans. from Jones (2006).
- On this, see Roller (2010), 48-9.
- On this sanctuary and its inscriptions, see J. Bingen (2007), 31-43.
- On this, see Hölbl (2001), 77-90.
- op. cit. 108.
- Though, contra Netflix, we can say something about her hair. I was rather confused when one of their talking heads posited that he liked to imagine Cleopatra with hair like his – his hair was very tightly curly and indeed that is how the Netflix documentary opted to do Cleopatra’s hair. We can see her hair on her coins and it is wavy, pulled together into a bun in Hellenistic style with short draping curls (loose, not tight) and a Hellenistic diadem. I think it is pretty fair to suppose this is how Cleopatra actually wore her hair – or at least, how she wanted people to think she wore her hair.
- The second edition of the Cambridge Ancient History; in this case CAH2 IX, 320-1.
- Literally “loves her fatherland”. To save myself some audio pain in the future, I often hear “philopator” pronounced with the emphasis on the second syllable (phi-LOPE-a-tor), but the accent is on the alpha, phi-lo-PA-tor, because that o is an omega, a long-vowel, so the accent can fall no further back.
- op. cit., 58-62 and 76-79.
- Bingen argues that in an Egyptian context, it would have been clear that Cleopatra’s patris was, in fact, Makedon – the most prestigious and useful possible patris to have in Ptolemaic Egypt.
August 4, 2026
QotD: Limits to tactical flexibility in Roman Republican armies
And therein lies the rub: just as the commander doesn’t have a multitude of points where he can intervene in the battle but rather a limited set of key decision points, so too he does not have an infinite set of potential orders he (or his subordinates) might give at any given moment. Instead, the commander (or his subordinates) has a McDonald’s menu worth of options he can employ. He can order anything on the menu and expect the unit to respond accordingly but he cannot order anything not on the menu, nor can he order an altered version of a menu item (no Big Mac without cheese; Sir, this is not a Wendy’s).
Now generals could and did use drilling in the early stages of the campaign to add one or two “menu items” to their drill so as to be prepared for this or that problem. Indeed the general preparing some sort of surprise of this sort and then deploying it against an enemy is a standard topos of effective ancient generalship (although in many cases we may doubt the purported gimmick was actually used (e.g. Dionysius 20.1.6, Dio 10.5)). But that is the thing: such extra “menu items” have to be prepared in advance and often well in advance of the battle. Caesar, for instance, anticipating war elephants drills his troops in anti-war-elephant tactics before the Battle of Thapsus (46 BC, Caes. Bel. Afr. 84). Alexander’s phalanx seems to have been drilled to respond to scythed chariots as well (Arr. Anab. 3.13.5-6).
That said the bulk of the menu was likely to remain constant. After all these “menu items”, while they have simple names are often actually fairly complex to do and train when you are dealing with units of dozens or hundreds. Turning back to our example of Company, Right Wheel! we can get a sense of the complexity of training that maneuver by the way they did it. The “right guide” (the man on the extreme right) adopts a set pace (nine inch steps) while the “left guide” moves at the normal regulated speed (note this requires training your soldiers not only to step in time but to step at standard lengths) and every soldier in the middle regulates his motion by maintaining elbow contact with the man to his side. That’s a good method to get the result, but it isn’t an obvious or simple method: the fellow training the company to do this needs to already know how to do this.
Now in the modern period with mass printing, you could print drill manuals that had all of the commands and instructions on how to do them, but in the pre-modern period that kind of mass printing isn’t an option (and our military texts almost never go into this kind of granular detail) so your officers have to learn to train the behavior by observing it and participating in it themselves. Which means in turn a great deal of your “package” is set before the army is even mustered at the beginning of the campaign season because you are relying on a package of culturally embedded knowledge about how armies ought to function. And if your culture doesn’t have that tradition, inventing it from scratch is going to be hard and take time; and by “time” I mean “decades”. Effective drill for volley fire (the “countermarch”), for instance, answered a need that had existed for most of the 16th century, but the embryonic ideas for it only started emerging in the 1570s, the actionable plan on how to do it in 1594 and it was only then finally deployed in a major battle in 1600 at Nieuwpoort.
However, the time available for that kind of training is limited. Remember that drill works by repeating the same set of actions over and over again until they become almost instinctive. That takes a lot of time and repetition! That both means it is going to be difficult to find space for new “menu items” (especially with green troops that still need to learn the basics) but also that synchronized discipline itself is expensive. Our ancient sources stress how Roman levels of synchronized discipline required diligent and continuous drilling to create and maintain.
Doing that in turn would demand keeping the soldiers in question under arms for training for long periods when they were not fighting. In the Middle Republic this seems to have been accomplished through rolling recruitment: Rome was essentially always at war and always raising armies so there were always a goodly number of veterans to train and “break in” the truly green new recruits in any given draft (though again – and I owe this notion to Michael Taylor – there are some signs that the less intense conflict of the mid-second-century leads to some dulling of the Roman edge). As we transition through the Late Republic into the Early Empire, that is replaced by the slow professionalization of the Roman army, culminating in the force of long-service professionals created by Augustus. Being mobilized in peacetime, professional soldiers could be drilled extensively (assuming their cultural values are going to let them drill you; also be wary here of assuming standing forces of “military settlers” are professional in this way).
But all of that is extraordinarily expensive. Keeping those soldiers under arms, after all, means that they need to be fed, clothed and paid. Even in a system like the Middle Roman Republic where soldiers provided their own kit out of their own wealth (because Roman soldiers in the Middle Republic were recruited from the modestly well-to-do smallholding class), by drafting those men you have removed them from that economic activity. The expense of doing this would have been in most cases enormous and of course those are resources that could be deployed to any number of other uses, including simply raising a larger army of less disciplined, well trained troops. For, say, a Greek polis given the choice between having 450 men under arms all year round and excellently trained or having 5,000 hoplites under arms only in the one month where there is a battle to fight the clear answer is the latter. A lot of things need to line up before producing armies with high levels of synchronized discipline makes sense: the state needs a lot of resources, year-round security challenges and a tactical system which is effective and demands this kind of investment. So while developing synchronized discipline offered an army a degree of greater control and flexibility, it was prohibitively expensive.
Bret Devereaux, “Collections: Total Generalship: Commanding Pre-Modern Armies, Part IIIa”, A Collection of Unmitigated Pedantry, 2022-06-17.
July 29, 2026
QotD: Asymmetric warfare
Bret Devereaux, “Miscellanea: A Very Short Glossary of Military Terminology”, A Collection of Unmitigated Pedantry, 2022-03-25.
July 23, 2026
QotD: The development of the Main Battle Tank (MBT) and Infantry Fighting Vehicle (IFV)
The result is the modern “main battle tank” (MBT), typically between 45 and 75 tons, with a single powerful direct-fire cannon as its primary armament. That’s not to say design differences between militaries drop away; there are still decisions to be made: the level of emphasis between speed, armor and firepower for instance. Another major decision is whether a tank is designed with a three-person turret (one of whom is a dedicated loader) or a two-person turret with an autoloader (that is, a machine that does the loading for you). It used to be that now I’d have to go on about why going autoloaders aren’t simply obviously the best but given the live demonstrations of some of the drawbacks of Soviet/Russian carousel autoloaders going on right now [in Ukraine] I’ll just say that both options have advantages and drawbacks and move on. So there is variety within the idea of an MBT, but by and large most countries field just one type of MBT (or one family of MBTs for countries which can’t afford to retire older systems) and that covers their tank needs.
But of course nothing could easy so just as tanks were getting easier to keep track of, APCs get much more complicated. Even before they were fully enclosed, APCs had generally carried some kind of defensive weapon (usually a heavy machine gun), but you weren’t supposed to fight with it if you could avoid it. Beginning in the 1950s, designers – first in West Germany – began experimenting with the idea that an APC could be armed to not merely deliver the infantry (“dismounts”) but to actually fight when they got there. The thing is, there are a lot of situations where an infantry formation might want more firepower but where a tank would be overkill: armored vehicles lighter than tanks, or enemies with lighter fortifications, for instance. West Germany pioneered the idea by putting a 20mm autocannon on their APC, the Schützenpanzer Lang HS.30, enabling it to put out some meaningful firepower, albeit much less than a tank.
The USSR cottoned on to the idea and liked it, resulting in the Soviet BMP-1 and the subsequent BMP family of what we call “infantry fighting vehicles” (IFV). Initially resistant to the idea (American doctrine was that APCs move troops and should not fight if they could help it), the United States, noticing the effectiveness of these sorts of platforms, got into the act with the Bradley Fighting Vehicle (these days the M2 Bradley). And IFVs, especially the Bradley, look like tanks. And indeed, they are sometimes called on to do the kind of armored reconnaissance and infantry support jobs that in a previous era would have gone to a light tank.
But they are also notably not tanks in several key ways. The primary job of an IFV is to move infantry and support them against lighter targets; some IFVs do maintain some kind of anti-tank capability, but it tends to be more limited (e.g. the Bradley’s pair of TOW AT missiles). The requirement that an IFV needs to be fairly fast but also needs to fit a bunch of dismounts means they can’t carry the firepower or the armor of a tank, often using lighter aluminum armor rather than steel or heavy composites. Finally, the tendency is for IFVs to mount lighter primarily armaments – again, space and weight concerns – like the Bradley’s 25mm M242 Bushmaster chain gun. The BMP-3 is probably the most heavily armed IFV out there with a low-velocity 100mm rifled gun and a 30mm autocannon, which is still a far cry from the firepower of the high-velocity 125mm main gun of a contemporary T-80 or the 120mm high velocity gun of a M1A1 Abrams. Instead, most IFVs would rely on anti-tank missile systems to engage tanks (although IFV autocannons can cause a tank serious problems if they can engage the rear or side armor, which is thinner).
Note that the IFV did not mean the end of the APC just as the APC didn’t bring the end of soft-skinned troop transports. In the US military, the Bradley serves alongside lighter-but-still-IFV Strykers, the venerable M113 APC along with soft-skinned HMMWV (Humvees; to be in theory replaced with the Oshkosh L-ATV) and trucks of various kinds and uses. Likewise, the Russian army has its BMPs (which are IFVs), and BTRs (which are APCs) as well as soft-skinned vehicles – or at least it did until, I am now being told, they were all towed away by Ukrainian farmers. The real Iron Harvest.
The final category worth noting here are variations on the concept of the MRAP (Mine-Resistant, Ambush-Protected): infantry transport vehicles designed to resist the sort of dangers one encounters in counter-insurgency operations, mostly mines, IEDs and small-arms ambushes. In essence, these vehicles are a purpose-built version of the up-armored Humvees designed to operate in a very particular combat environment (insurgencies and particularly urban insurgencies). These sort of vehicles have gotten more visibility in the United States because many of them have been shifted into the hands of law enforcement, where their utility is questionable, to say the least.
Bret Devereaux, “Collections: When is a ‘Tank’ Not a Tank?”, A Collection of Unmitigated Pedantry, 2022-05-06.
July 17, 2026
QotD: Catapults in pre-gunpowder armies
The very oldest catapults, first invented by the Greeks were tension catapults (the gastrophetes and oxybeles), which functioned like large bows, with a bow-staff being bent backwards to store and then release the launching energy. This sort of design, common in pop-cultural depictions of catapults, is actually quite limited as with the materials available, there is a real limit to how much energy can be stored via tension. Fortunately for the Greeks, by the early fourth century, they had developed a better method.
Instead, the Greeks, Macedonians and Romans began using torsion catapults (where the energy is stored in wound-up sinews like a spring). While the devices used in field battles (and for city defense) were often smaller, arrow-launching devices, siege catapults could be very large; the standard engine for the purpose could fling a 1 talent stone (26.2kg) about 400m (though effectiveness was far higher if you could get closer to the wall, which as we’ll see will be a trend for most of this post); much larger engines did exist as well. That said, Roman catapults were mostly not for collapsing walls but for destroying towers and suppressing defenders in order to aid in escalade (usually by mole, rather than ladders or towers, though the Romans used those too).
And here once again the distinction between the “big army siege package” and the “small army siege package” matters quite a bit. Roman torsion artillery was complex, expensive and required lots of technical skill, and so sees far diminished use in the early Middle Ages where that technical skill is hard to come by. Vespasian, we are told, brought 160 torsion catapults to besiege Jotapata in 67 (Josephus BJ 3.166) while Titus brings a stunning 340 to besiege Jerusalem in 70 (Josephus BJ 5.356). By contrast, the construction of a single catapult is often a major event in a medieval siege (see Rogers, op. cit. 121-3 for some examples) and while later medieval catapults were often more powerful than the earlier Roman torsion devices, they were not that much more powerful.
Consequently, Hellenistic and Roman fortifications (especially city walls, like the Theodosian Walls we discussed last time) were designed with massed catapults in mind. As noted, the multiple walls ensured that the main curtain wall, the inner wall, was extremely difficult to target with catapults or indeed any kind of artillery: even if you knocked down the low wall and the outer wall, their rubble would mostly block shots at the base of the inner wall. Meanwhile, the inner wall was built to be practically immune to catapult fire anyway: up to 6m thick without any internal passages (the outer wall was much thinner, only 2m). That was more than enough to render the walls effectively immune to anything catapults can do; the walls in many places still stood up to Ottoman cannon in 1453. Finally, ancient city defenses were built assuming they’d often have their own stone and arrow throwing torsion artillery set up on the towers to return “counter-battery” fire. Not every city had the “complete package” that Constantinople, as the imperial capital head, of course, but some mix of thick walls, low out-walls and catapults designed for counter-battery fire were fairly standard defensive arrangements for Roman cities that could afford them and felt sufficiently threatened to invest the resources.
As we move into the Middle Ages, two paradoxical things happen. On the one hand, the ability for societies in Europe to deploy large numbers of finicky, high-tech torsion artillery decreases dramatically (and the machines that we do see tend to be the simpler, less accurate single-armed variety, what the Romans called the onager or “wild ass” because it kicked like one when it fired). On the other hand, by the sixth century, we start to see a clever new design of catapult, the traction trebuchet.
Originating in China in the 4th century BC, the traction catapult used muscle power directly to swing a long pole around a central frame. In terms of engineering complexity, it was a simpler device, and could be scaled up quite large so long as one could add more pullers (around 100 seems to have been normal for a large engine), but the range and power it offered as a result of the mechanical advantage offered by the long throwing arm were considerable. Given the number of pullers required, it is little surprise these were generally only used in small numbers in medieval Europe (again, often in reports it is merely a single device, described as a mangonel or a fenevol), but on the other hand, as I understand the physics, the range and striking power had the potential to be superior to a torsion catapult. Nevertheless, if we look at the kinds of fortifications emerging during this period, it certainly seems like in Europe, the concern that artillery might produce a breach in the wall (as opposed to merely degrading towers and the wall-walk) was fairly low.
Just to throw down a note here because we’ll come back to it, it is striking that while the small numbers of traction trebuchets in Europe seem to have represented a decline in the “catapult threat” to walls (recall last week’s contrast between castle walls and the much older Theodosian Walls), that was not the case in China, where walls continued to be made very thick – a design quirk that will matter quite a lot in a moment. I am not an expert on ancient and medieval Chinese siege tactics, alas, but my brief encounters with accounts of them often seem to describe traction catapults used en masse, in dozens or even hundreds, much more the way that the Romans used massed siege artillery. Likewise, Michael Fulton (Artillery in the Era of the Crusades (2018)) notes nearly a hundred Mamluk trebuchets (a mix of counter-weight and traction) at the Siege of Acre (1291); my sense is that such large siege trains were very rare within Europe. Presumably the ability to deploy so many engines was a consequence of greater state capacity in China and the Near East during this period as compared to fragmented, decentralized medieval Europe.
The late 12th century sees a major variation on the trebuchet design: the use of a counter-weight, instead of traction to provide the force; this innovation seems to have emerged in the West broadly defined, though it isn’t clear if that means in Europe or the Middle East (in any event both Christian and Muslim armies start using them at almost exactly the same time). This allows for much more energy to put into the shot, as the counter-weight can be very heavy and only slowly winched into place, allowing the work crew to spend more time “storing” energy in the counter-weight than they could with the quick pull of a traction trebuchet. Larger counter-weight trebuchets could also make use of animals to provide the power, or large wheels to make it easier to raise the counter-weight. The upper-limits on the size of projectiles were very high: Warwolf is thought to be the largest such trebuchet known, and threw a nearly 300lbs shot. That said, while counter-weight trebuchets hit harder (but fired slower), in function they do not seem to have been meaningfully different from traction trebuchets; they were used the same way in sieges.
What’s really striking is not the vast impact of catapults, but the muted impact of catapults. The counter-weight trebuchet was clearly good: the innovation makes its way all the way back to China, carried by the Mongols who presumably picked it up in the Middle East (ironically moving the opposite direction but at the same time as gunpowder, suggesting that at this point in the 13th century the two technologies were not considered mutually exclusive). Castle design does respond to catapults, but only in relatively modest ways: walls get somewhat thicker, but as Fulton (op cit.) notes, only by about half a meter or so (leaving even the newly thickened medieval castle walls somewhat thinner than the best old Roman defenses). In at least some areas, towers and keeps become more frequently rounded in shape, to resist catapult fire.
Certainly it was possible for catapults to open breaches in weaker walls to enable assault. The aforementioned Warwolf opened large breaches in the stone walls of Stirling Castle in 1304. But I note both Rogers (op. cit.) and Fulton (op. cit.) seem to confirm that while true breaches from trebuchets could happen, it was far more common that walls resisted trebuchet strikes and that the real work of the machines was degrading the wall defenses by striking off battlements and smashing towers, in order to enable escalade. Which is little surprise: that’s precisely what the Romans used catapults for too. While there is still some argument about the degree to which the counter-weight trebuchet was a revolutionary military technology, on the balance, the siege playbook changed only modestly to accommodate it, and castle design likewise shifted only in degrees.
And then Charles VIII of France (r. 1483-1498) decided to take a holiday on the Bay of Naples.
Bret Devereaux, “Collections: Fortification, Part IV: French Guns and Italian Lines”, A Collection of Unmitigated Pedantry, 2021-12-17.
July 11, 2026
QotD: Could airpower have broken the trench stalemate on the western front in WW1?
There are two directions to take airpower here: tactical and strategic. One wasn’t ready then (but would be by WWII), the other still hasn’t managed to accomplish its stated objectives yet, but continues to over-promise and under-deliver results.
Let’s deal with tactical airpower first. The first function aircraft were put to in WWI was reconnaissance. In 1914, that might mean locating the enemy in a fast-moving battlefield, but as soon as the trench stalemate set in, reconnaissance mostly meant identifying enemy buildups along the line and – still more importantly – serving as spotters for artillery. It wasn’t a huge cognitive leap to go from having aircraft which identified targets for the artillery to thinking that the aircraft could be the artillery. But as with tanks, the technical limitations of the platforms in use meant that actually meaningful close air support was still two decades away when the war ended. The rapid development of aircraft in these early days means that there is a truly bewildering array of aircraft designs in use during the war, but the Farman F.50 is a good sample for what the most advanced bombers in common use looked like towards the war’s end. It carried a maximum of eight 44kg bombs (352kg) under the wings, which were dropped unguided. With a maximum speed of less than 100mph and a service ceiling under 5000m, it was also an extremely vulnerable platform: fragile, slow and with a relatively low flight ceiling. The French mainly used bombers at night for this reason.
But how much airpower does it take to really move a division out of position? In 1944, at the start of Operation Cobra as part of the Normandy breakout, it was necessary for US forces to move the powerful armored division Panzer Lehr out of its prepared positions outside of St. Lo. Over the course of an hour and a half, the U.S. Eighth Air Force hit Panzer Lehr with approximately three thousand aircraft, including 1,800 heavy bombers (each of which might have had bomb-loads of c. 2-3,500kg; the attack would have been the equivalent of about 13,000 Farman F.50s (of which only a hundred or so were built!)). By this point, even medium bombers carried bomb loads in the thousands of pounds, like the B-25 Mitchell medium bomber, with a bomb load of 3000lbs (1360kg). This was followed by a hurricane artillery barrage! Despite this almost absurdly awesome amount of firepower (which, to be clear, inflicted tremendous damage; by the end of Operation Cobra, Panzer Lehr – the heaviest and most powerful Panzer division in the west – had effectively ceased to exist), Panzer Lehr, badly weakened was still very capable of resisting and had to be pushed out of position by ground attack over the next three days.
Needless to say, nothing on offer in 1918 or for a decade or more after, was prepared to offer that kind of offensive potential from the air. That kind of assault would have required many thousands of aircraft with capabilities far exceeding what even the best late-war WWI bombers could do. Once again, while close air support doctrine was developed with one eye on the trench stalemate and the role airpower could play in facilitating a breakthrough and restoring maneuver (either by blasting the breakthrough or – as in Soviet Deep Battle doctrine – engaging enemy rear echelon units to bog down reinforcements). But the technology wasn’t anywhere near the decisive point by 1918. Instead, the most important thing aircraft could do was spot for the artillery, which is mostly what aircraft continued to do, even in late 1918.
But that’s tactical bombing against military targets. What about strategic bombing against civilian targets?
The first efforts at strategic bombing were made in WWI, though once again the technology wasn’t ready. The range for fixed-wing aircraft was still very limited; the aforementioned Farman F.50 had a range of only 420km, nowhere near enough to really bring entire countries under the threat of bombing. Dirigibles – zeppelins – could manage much longer ranges and the Germans did attempt to bomb British cities with them starting in 1915. The problem was that once aircraft powerful enough to climb to the zeppelin’s altitude were developed, the slow and fragile zeppelins were sitting ducks: lighter than air airships could hardly be armored, after all. Moreover, the bomb loads of zeppelins had always been far too low to make effective strategic bombing possible beyond the initial shock of it.
What no one could have known in WWI was not merely that the technology for effective conventional strategic bombing wasn’t ready, but that it would probably never be ready. Interwar air-power theorists, seeing the potential of strategic airpower to bypass the trench stalemate by flying over it began to try to work out how this would be done. Giulio Douhet (1869-1930) argued that future wars would be fought and won in the air, with fleets of bombers using high explosives and chemical weapons to massacre enemy civilian centers, until civilians forced their governments to surrender. Douhet was not alone; his vision of airpower was shared, for instance, by the “father of the RAF”, Hugh Trenchard (1873-1956).
This concept, “morale bombing” as it is sometimes called, probably deserves its own post discussing its failures. But in brief, the concept was tested, with far larger amounts of bombs than Douhet or any other interwar theorist could have ever dreamed of, during WWII. The argument by air theorists that high altitude bombers could not be stopped was proved false when the British did exactly this, stopping German bombers over Britain in 1940. Moreover, terror bombing against civilian targets in Britain didn’t lead to surrender, but hardened resolve. Likewise, “morale” bombing against German targets by the allies didn’t lead to surrender, but hardened resolve. Later efforts to demoralize the North Vietnamese through a American bombing campaign in the Vietnam War didn’t lead to surrender, but hardened resolve. More recent efforts to demoralize or destroy terrorists and the Taliban through the use of airpower hasn’t lead to surrender, but rather hardened resolve. Likewise, efforts by the Syrian Regime to defeat various opposition groups in Syria through the use of chemical weapon-based terror bombing didn’t lead to surrender (siege-and-starve tactics did), but hardened resolve.
It turns out the fundamental premise of the entire idea of morale bombing – that being bombed will make people want to stop fighting – was flawed. Morale bombing has been, depending on how hard you squint at the US air campaign over Japan in WWII (including the use of nuclear weapons) successful either once (out of many attempts) or never. In most cases, the sustained bombing of civilian centers has been shown to increase a population’s willingness to resist, making the strategy worse than useless.
The case for strategic bombing against industrial targets is marginally better, but only marginally. While airpower advocates, particularly in the United States promised throughout WWII that bombing campaigns against German industry could lead to the collapse of the German war machine, in the end many historians posit that the real achievement of the campaign was to lure the Luftwaffe into the air where it could be destroyed, thus denying the German army of air cover and close air support, particularly on the Eastern Front. Some diminution of German industrial capabilities was accomplished (though it is not clear that this ever approached the vast resources poured into producing the large numbers of extremely expensive bombers used to do it, though the allies had such an industrial advantage over Germany, forcing the Germans to fight in expensive ways in the sky was a winning trade anyway), but the collapse of German industry never happened. As Richard Overy notes, German industrial output continued to rise during strategic bombing and only began to fall as a result of the loss of territory on the ground. Needless to say, “strategic bombing can sucker the enemy into wasting their close air support” was not the result that airpower advocates had promised, nor could it have broken the stalemate.
I don’t want to oversimplify the continued debate over the efficacy of strategic airpower here too much so let’s just say that the jury is still very much out as to if strategic airpower works even with modern technology; it certainly wouldn’t have worked with WWI era technology.
Bret Devereaux, “Collections: No Man’s Land, Part II: Breaking the Stalemate”, A Collection of Unmitigated Pedantry, 2021-09-24.
July 5, 2026
QotD: Battlefield communication in the pre-modern world
The first option, of course, is shouting. The problem here is obvious: how is the commander’s command to be heard? Interestingly, there has been a fair bit of research by ancient historians looking at the question of how many people can possibly hear a short address unaided by modern loudspeakers and the like; figures vary but generally a few thousand if they are reasonably compact and quiet. That might work for a general’s pre-battle speech, delivered before the army advances, but it will not do for an army that is already in motion, much less once the chaos of battle has begun. Thousands of men marching (let alone fighting!) are noisy!
The modern solution to this problem is radio, but of course that’s hardly available to our pre-modern commander. Instead, to judge by films, the mind quickly jumps to signal flags. I am reminded of Braveheart (1995)’s rendition of the Battle of Falkirk, where Edward I uses signal flags to order his archers forward. HBO’s Rome also does this in its version of the Battle of Philippi, with flags being jostled and then pointed forward to signal the advance. Unfortunately, signal flags – as distinct from unit flags (which we’ll come back to in a moment) – have a few key problems, the most notable of which is that no one will be looking at them: after all the army is advancing, the soldiers are looking forward but signal flags (again, as opposed to unit flags) are going to be behind them, not placed out in the middle of No Man’s Land between the armies. As a result, signal flags are useful for sending information long distances (in a chain of stations or operators), for instance from one commander at distance to another, but not in battle; operational, rather than tactical tools. In practice, the use of signal flags like this is confined to the modern era; the first successful “optical telegraphs” (as iterations on things like smoke signals and fire relays) date to the late 18th century.
Unit flags – a banner or other big, obvious symbol (like a statue of an eagle on a stick) – are more useful. These can be positioned at the front of a unit, typically at its center. If it advances, then the soldiers in the unit also know to advance, following the standard they can see (because it is elevated, large and visible) even if they cannot hear the orders. There are two complications here though: first, the unit banner or flag is a relatively late innovation in antiquity, really only coming into its own with the Romans. The Achaemenids may have used some kinds of ensigns or standards, but the Greeks do not seem to have done. Instead our first really good documentation of something like a battle flag comes from the Romans: each legion had a signa (eventually standardized to the legionary eagle, the aquila), which was a shiny metal statue mounted on a pole so it could be easily seen. Units of the legion broken off to do other things might instead follow a less impressive cloth banner, a vexillum, by which such detachments became known as vexillationes. But the broader problem is that of course your general may not be particularly close to your flags (or other standards) which are generally at the front-center of each component unit of your army. The flags may allow a subordinate officer to “drive” the unit over the battlefield – and that’s good – but it doesn’t let the general tell that officer what to do.
A better option is music, but once again development seems to come fairly late in antiquity. Greek hoplites seem to have advanced to the music of the aulos, a double-reeded flute-like instrument; given the limitations of the instrument it is generally assumed it was used to keep time (so everyone marched in step) not transmit orders. Once again, a more complex system of musical signalling seems to come with the Romans, at least as detailed by Vegetius. Vegetius (2.22) notes three different kinds of horn instruments used by a legion: the tubicen was used to sound charge and retreat, the cornicen regulated the movement of the signa (so “advance” or “halt”), while the buccina was used mostly for camp signals: sounding watches or assemblies. It’s a system that is akin to later bugle calls, but note that the orders it can give are limited to a relative handful of prearranged signals: advance, halt, charge, retreat, assemble, change shift and so on.
The attentive reader here may have already noticed how developed Roman command and control is and may suspect that ties in with the Romans having a more “command” oriented culture of generalship; if so you are ahead of the game!
Of course if those instruments are sounding on a per-unit basis (and they are) that means you still have the problem of getting the order from the general to the instruments for the unit in question. And fundamentally here, the technology is – as I tell my students – man-on-horse. The particular fellow on the horse may be a dedicated messenger (if your military organization has those) or a subordinate officer or it may be the general himself.
But it is important to note now the limitations of this sort of system and we can use what we know of the Roman command and control system (as noted, one of the more developed of such systems prior to gunpowder) to get a sense of them. Let’s say the general realizes there is a problem on his flank and he needs a unit (probably here we’re talking a cohort or a maniple, not a legion) to change what it is doing. First off, the order needs to get within shouting range of the unit’s commander (in this case a senior centurion). The general can either go themselves or send a messenger; both options have their downsides. If the general goes himself he is essentially removing himself from observing or commanding the rest of the battle, but a common problem with sending a junior subordinate is that the unit commander may not respect or feel the need to obey that subordinate (written orders can help with this, but now we’re bringing in questions of literacy). Of course both a messenger or a general in transit may also well be killed, which will prevent the order from being received!
In either case, the message is going to move at galloping speed, which is around 40km/h, meaning that it may take several minutes for the general or messenger to navigate to the spot. That doesn’t sound so bad, but battles with contact weapons do not typically go for hours and hours; Pydna (168) was, as noted last week, decided in about an hour total! Of course a battle might be longer (or shorter!) than this, though much of that extra time is likely pre-battle skirmishing – the actual direct press of infantry formations in shock rarely lasts long because of the terror of it (and to a lesser extent its lethality; we’ll return to the balance of terror and lethality next time). Imagine if you were playing a Total War game and your input delay was, say, five minutes long in a battle that might only last an hour or two.
But of course galloping time isn’t the end of it. The message now has to be conveyed to the unit. In the Roman system, that means the messenger needs to find the appropriate centurion, explain the order to him and then ideally that fellow will then signal the instruments and signa to act accordingly – but even then, those instruments and signa only have a handful of prearranged signals available. Anything more complicated will need to be shouted down the line the old fashioned way (as we know, for instance, the Spartans did for lack of almost any of the rest of this apparatus of command, Xen. Lac. 13.9). Needless to say that means that giving any complex order to a unit already engaged or about to be engaged is going to mean starting by signalling retreat and then attempting to regroup the unit; regrouping an already retreating unit is one of the most difficult tasks on a battlefield and is rarely performed successfully in an unplanned fashion (even in a planned fashion it goes wrong as often as it goes right).
(This is, by the by, why reserves are so important. An unengaged unit hanging behind the lines can be given new orders far, far easier than a unit that is already engaged or about to be. And indeed, those familiar with the Roman system of fighting with its three lines of heavy infantry will note that it is a system heavy on reserves. Indeed, the manipular legion essentially assumes it will be necessary to retreat and regroup the first line of heavy infantry (the hastati) behind the second (the principes) and plans and drills for that. Note how the Roman command culture, the Roman fighting method and the actual apparatus of messengers, signa, instruments and junior officers all align here – that’s common because these sort of institutions tend to co-evolve.
By contrast we may compare a Greek hoplite army in the Classical Period. It has no battle flags or ensigns and the general is expected to fight on foot. In the past I’ve described the resulting phalanx as an “unguided missile” and this is a big reason why. That’s not to say hoplite generals never exerted command on the battlefield – better generals might keep a reserve to be rushed to important points (as Pagondas does at the Battle of Delium in 424 BC). But for the most part, once a hoplite general formed up the army and hit “go”, they had very little control over the army.
Bret Devereaux, “Collections: Total Generalship: Commanding Pre-Modern Armies, Part II: Commands”, A Collection of Unmitigated Pedantry, 2022-06-03.
June 28, 2026
QotD: Getting cloth to market in the ancient and medieval world
It should thus be of little surprise that regions involved in major textile production for export were often concentrated either on coasts or on rivers that were navigable to the sea (one may map the regions Pliny lists as major wool and linen exporters to find that they are all accessible by sea). While the sheep themselves may be grazed part of the year up in the uplands far from the coast, one of the great advantages of transhumance is that the sheep may transport themselves under the care of their shepherds to villages and lower pastures not too far from coastal towns which may serve as centers of textile production and major points of sale.
Now those transport costs become less and less significant the more valuable the goods being transported are. For a bulk good like grain (or common wool), transport may represent a majority of the costs. But if one is shipping something extremely valuable (particularly valuable per unit weight), the cost of acquisition at the source (and the profits of final sale) are much larger relative to the transport costs and less efficient methods of transportation become useful, thus the viability of silk and other expensive luxury goods being transported overland across Eurasia on the famous Silk Road.
Very high value fabrics didn’t need to come from so far afield though. In the Roman world, the province of Asia (corresponding roughly to western Turkey today) had several notable centers of production for particularly high valued textiles (on this, see I. Benda-Weber, “Textile Production Centers, Products and Merchants in the Roman Province of Asia” in Gleba and Pásztókai-Szeöke, op. cit.). Thyateira’s guild of purple-dyers (the πορφυροβάφοι) seem to have had trade contacts for their wares – wool dyed Tyrian purple via the murex snail – all over the province as well as in Macedonia and Italy. Weavers in the region were also known for producing fabrics with complex woven patterns and Miletus, one of the major ports in the region, had as noted the reputation for producing the best dyed wool in the Mediterranean. Such fabrics were highly valued and we find evidence that such fabrics were bought not merely by the Roman elite, but also made overland as far as Persia where such wares were valued at the Achaemenid (550-330 BC) court.
Neverthless, not all fabrics moving through trade in antiquity or the middle ages were rare or high value fabrics. As Jinyu Liu notes in a study of inscriptions relating to the textile trade, “coarse wool and wool of medium quality, and products made of these non-luxury wools dominated the market” in the Roman Empire, often being “pulled” through trade towards both large population centers in the interior of the empire and towards the Roman armies in the frontier provinces, both of which must have outstripped local production in their demand for textiles (Liu, “Trade, Traders and Guilds (?) in Textiles” in Gleba and Pásztókai-Szeöke, op. cit.). This trade included not just fabrics but also ready-made products like garments or blankets which must have been aimed at fairly modest people, neither the very poor (who couldn’t afford them) nor the wealthy (who wouldn’t have been caught dead in “ready-made” one-size-fits-no-one clothing), but rather the middling urban workers and common soldiers (and perhaps small farmers, though we might assume their households would produce most of their own textiles in the countryside where wool and flax, being agricultural and pastoral products, might be more available).
In Medieval Europe, just as in the ancient world, the centers of textile trading tended to follow the water as it made transport easier. England was a major wool-producing center in the high and later Middle Ages (and into the Early Modern period), with J.S. Lee (op. cit., 9) estimating production per capita exploding from around 1.3 pounds per person in the early 1300s to 7 pounds by the 1550s as the textile production system in England reoriented towards export. Wool products, produced in towns mostly in towns that were nearly coastal or had river-access flowed down by coastal trade and up the Thames to London to either be sold and used there or to be further exported to the dyers and fabric markets of the Low Countries (where fabrics could use the Rhine to travel further into the continent) or to be bought by the merchants of the Hanseatic League and so head into the Baltic.
Bret Devereaux, “Collections: Clothing, How Did They Make It? Part IVb: Cloth Money”, A Collection of Unmitigated Pedantry, 2021-04-09.
June 22, 2026
QotD: When the US switched to the All-Volunteer Force in 1973
This was a huge change. For such a major change, I find that it draws surprisingly little attention. The 50th anniversary of the AVF passed with relatively little fanfare in 2023. I’ve mentioned For the Common Defense (1984, 1994, 2012) as the dominant textbook for introductory American military history: the shift to the All-Volunteer Force is dealt with in a single page (page 568, for the curious). The textbook I’ve seen most recently used for US Naval history (and which I used), J.C. Bradford and J. F. Bradford, America, Sea Power and the World (2016, 2023), doesn’t even give it that much: the shift is discussed in a single paragraph on page 351 (308 in the 2016 edition).1
The likely impacts of the shift to an AVF were studied prior to implementation in the Gates Commission, a report that had a preordained conclusion – it was convened to provide Nixon the cover to do the thing (end the draft) he had promised to do already in his campaign – and which honestly I find disappointing in its approach, which is mostly “happy talk” designed to justify what Nixon had already decided to do. It is striking to me, for instance, that the Gates Commission did not include a single historian to perhaps discuss how the shift towards fully professional militaries had gone for republics in the past. Instead, the focus is on the economics of the shift, with fairly blithe assertions that the civil-military relationship would remain unchanged despite the fairly obvious implausibility of that given the shift from “everyone serves” to “only a small portion of society serves”.2
As I’ve noted elsewhere, the Romans also seem to have thought that they could professionalize their army without reducing its ability to scale up in an emergency or altering the civil-military relationship and for quite a few decades that more or less worked, while the old norms held. But as those old norms decayed, the institution increasingly became what you’d expect from its institutional structure: a permanent political faction, advocating for its own interests, often with violence, to the point that the emperor Septimius Severus’ advice to his sons as he lay dying in 211 was, “Be harmonious, enrich the soldiers, and scorn all other men”,3 a fairly open admission that the soldiery was not just a political constituency, but the most important one. It took time for those norms to shift, but when one is building or rebuilding institutions, the long-term is the term that matters.
I do not think necessarily that this is the direction the All-Volunteer Force must go. It has two and a half centuries of strong norms pushing it away from this direction. But careful maintenance of the civ-mil bargain is made all the more necessary when the military is effectively fully professional. For my own part, all cards on the table, while I greatly value the service of the United States’ military personnel (there’s that third part of the bargain!) and think they serve honorably, I am quite skeptical of the long-term implications of the All-Volunteer Force. Its creators assumed that fully professionalizing the military would not impact the civil-military relationship and that it would always be possible to shift back to a mass-conscript army in the event of a major war, but historical examples suggest it is not so easy.
But the All-Volunteer Force is not the direction from which I see now the principal threat to the civ-mil bargain.
Bret Devereaux, “Collections: The American Civil-Military Relationship”, A Collection of Unmitigated Pedantry, 2025-07-04.
- In that book’s defense, the Navy has a really big set of reforms associated broadly with CNO Elmo Zumwalt that happen at basically the same time and are connected and it opts to focus on those. I will note that the position of the paragraph has changed because the updated 2023 version of the book has opted to grapple more extensively and more successfully with this period as one of increasing diversity in the navy, with a chapter by Kristy N. Kamarck on that specific topic. It is a marked improvement over the first edition, though I think both FtCD and the Bradford and Bradford remain too hagiographic, too willing to sweep the military’s problems under the rug and only comment on military diversity when they can tell the story as a happy tale of progress.
- Especially as that small portion tends to be concentrated, a thing the Commission essentially refuses to consider as a first principle of their analysis; they assume cheerfully that the AVF will naturally continue to reflect a cross-section of the United States. In some ways that is true, but in other ways it is very much not – there certainly are “military families”, where service tends to “run in the family” in the United States now – and the emergence of those patterns would have been a pretty obvious thing to expect, given that the same trend is extremely visible in the Roman army of the early empire.
- Dio 77.16
June 16, 2026
QotD: Nitpicking the field fortifications in Gladiator (2000)
What we are shown is pretty clearly a prepared defense on a hillside, with a series of raised terraces, with a mix of abatis (sharpened wood obstacles, often crudely cut wood stakes set in an X pattern) and mantlets, with gaps in those defenses to allow units to move and a whole bunch of catapults positioned up on the hill. The terraces make for a layered, multi-stage fighting position at each level. On the one hand, the Romans were hardly averse to field fortifications and one wonders again if this set was a product of someone with an active imagination looking at the Column of Trajan [Wiki], which features a lot of scenes of Roman soldiers cutting trees and building bridges, roads and forts.
The problem isn’t that there are field fortifications, it is everything else about them: the style of field fortification, their position, layout and use. As we’ve noted before, Roman armies on campaign built fortified marching camps nightly, so we would expect Maximus’ army to have such a camp, but as we’ve discussed even more so, one of the classic, famous features of Roman armies is that they build the same layout of camp wherever they go, the famous Roman “playing card” forts, generally built on flat, open ground (rather than hillsides). That defense would not look like this, instead consisting of a ditch (the fossa) behind which would have been a earthwork rampart (the agger) topped with a wooden palisade (the vallum); thus rather than successive layers, you’d have a single clear fighting position (the vallum) on a mount with the ditch directly in front of it. And that would be a continuous line, with just four gates (at the center of each side), rather than this kind of checkerboard pattern of fortifications, because of course the purpose of this defense was to prohibit entry. Moreover, the line of field fortifications we see are not part of, nor connected to, a marching camp: it is simply a line of fortifications on the side of the hill with nothing on the flanks, rather than the distinctive “playing card shape”. We don’t see the camp sitting behind it either.
But the really immediate problem is that Maximus’ army has formed up within his troops strung through the field fortifications, with legionary soldiers mostly in front of them (but some are behind them) and the archers in between the stakes and mantlets. This may seem like a sensible way to form up a defense, but it is not the Roman way. Maximus is very intentionally “offering battle”, – inviting his opponent into an open field engagement. The way a Roman army did this was invariably forming up on the flat, open, unfortified ground in front of the camp, toward the enemy, signalling that they would fight in the open, outside of their walls (as Maximus does indeed intend to do).
So what we ought to see is Maximus’ army formed up outside in the open field, with the camp likely visible some distance behind them. That camp would be protected by very different fortifications: you’d be able to make out its “playing-card” shape, with watch-towers on the corners and the raised vallum running the exterior and the relatively neat grid of tents in the interior.
Finally, before we get to the battle plan, I want to note one more oddity here, which is the battlefield itself. The battlefield is a muddy field, which it looks to have been recently clear-cut, otherwise surrounded by dense forest. Of course part of the reason is that this is Bourne Wood, a coniferous tree plantation (and frequent filming location) in Surrey, England (which is why the trees are all the same species, so neatly spaced out) rather than the edge of an old-growth forest somewhere in southern Germany.
But the thing is, the Marcomanni, Quadi and other Germanic-language speaking peoples were an agrarian society, same as the Romans: their villages were surrounded by farm and pastureland. Of course a lot of the forest – old-growth forest, rather than tree-farms as here – remained, but if a Roman army wanted a flat, open space to offer battle in, they needn’t have cleared it themselves (and indeed probably couldn’t, at least not in the time frame they’d have to prepare for a pitched battle), but could simply march to the nearest village with its patches of farmland. Getting a Roman army to fight in dense, old-growth forest, after all, famously required clever ambushes, as a Teutoberg Forest (modern Kalkriese) in 9 AD. And if the enemy didn’t want to fight in the open, Roman armies were perfectly happy to burn villages and pillage crops as the standard way of attempting to force an enemy to accept an offer to battle or else vacate the area.
Bret Devereaux, “Collections: Nitpicking Gladiator’s Iconic Opening Battle, Part I”, A Collection of Unmitigated Pedantry, 2025-06-06.
June 10, 2026
QotD: Tiberius Gracchus, Tribune of the Plebs
And here it is worth noting that kind of government the Romans had to understand the response. The Roman Republic had written laws but no written constitution – instead, the rules for office holding, for conducting the business of the Senate, for running the assemblies and so on were all customary: the Romans governed themselves in accordance with what they called the mos maiorum, “the custom of the ancestors”. In a sense then, certain practices, if practiced long enough, became a sort of law-of-tradition to themselves and of course one of those customs – practiced at this point for, at minimum around 150 years – was the continual leasing of large amounts of ager publicus to the point that the leases were treated as a form of ownership: people used that land as security for loans, they built houses on it, they buried their parents on it and so on. Because the leases were presumptively renewable and had been for decades if not centuries, under the mos maiorum, the holders of ager publicus had long considered the land theirs. And of course the upset parties are rich and powerful, so their opposition was significant and meaningful, politically.
In brief, the way this plays out is that while Tiberius Gracchus does have significant popular support for his motion (Plut. Ti. Gracch. 9.1), much of the elite are opposed. He draws up a quite conciliatory version of the law, which proposes to compensate the holders of large amounts of ager publicus for their lost leasing rights and to then give them the remainder of their leased land (so they needn’t fear a second lex agraria and a third and a fourth and so on), but according to Plutarch in the face of continued elite opposition, shifts back to a less conciliatory version of the law (Plut. Ti. Gracch. 10.3). The resistance to his law centers on another tribune, Marcus Octavius, himself a large holder of public lands, who plans to veto the law and uses his own powers as a tribune to disrupt the process (along with some fairly clear shenanigans by some of the wealthy, like trying to hide the voting urns to prevent a vote on the law and so on).
Now there are a few things to note at this juncture in the story. First, there being ten tribunes, it must never have been very hard to find a tribute willing to gum up the passage of a given law, but that, traditionally, this was a tactic of delay, rather than a hard-stop the way Octavius is using it. At the same time, with real public momentum to make this law happen, one could easily imagine simply waiting Octavius out – he only has one year in office. Except. Except that, remember, Tiberius Gracchus needs a big victory in his tribunate to get his political career [back] on track, a consideration that was clearly significant (thus the reason we’re informed of his quaestorship; we usually don’t know much about even very significant figures’ time in junior offices!). That consideration, I think, serves as important context for Tiberius’ decision to escalate every time he encounters resistance: he cannot afford to simply be the prelude to someone else passing this law: he needs to pass it himself.
The normal method for “deconflicting” two magistrates with opposing vetoes like this was to go to the Senate, which Tiberius Gracchus, hoping his influential supporters would carry the day, did. Instead, according to Plutarch (Ti. Gracch. 11.2) the Senate was merely no help, whereas Appian (BCiv 1.12) describes the Senate as openly upbraiding Tiberius, a strong negative response. Now under the mos maiorum, that would be the end of it: the authority of the Senate (the auctoritas senatus) ought to be so intense that when the Senate speaks in one voice and says, “not right now” then you desist. Remember that in the Roman conception, the Republic is a partnership of sorts between the Senate and the People (the S and the P in SPQR), rather than a situation in which the Senate is purely subordinate to the popular will: if the Senate is strongly opposed, that is supposed to be a veto point that is respected.
But remember: Tiberius Gracchus cannot, politically, desist. He must push through because his political career requires a victory this year. Note that the cause does not require a victory in 133; there is nothing to stop another tribune in 132 from trying to advance the same bill or a more limited or different version of it. But Tiberius Gracchus’ career absolutely requires success in 133. So instead of desisting, he escalates.
He now breaks clearly with the mos maiorum and plans to take his law directly to the people against the advice of the Senate. Octavius is obviously a problem – he’ll veto anything Tiberius Gracchus tries to do – so Tiberius Gracchus introduces a law to depose Octavius from office. The Roman Republic doesn’t have anything like impeachment, there is no framework to remove someone from office. Instead, the way the Republic works is that all of the offices are held for short duration (one year) and while tribunes and office holders with imperium are immune from prosecution while in office, they can be prosecuted the moment they leave office for any crimes they committed. There is no framework for booting out a tribune like this; the remedy in the customary Roman system is to make sure the next year you elect tribunes who support the idea and try to pass it then. But that remedy doesn’t work for Tiberius Gracchus.
So Tiberius Gracchus passes the law deposing Octavius and then has him dragged from the speaker’s platform (the rostra) and now we have a problem. Because of course Octavius’ supporters are going to view this law itself as illegal and invalid: tribunes are, you will recall sacrosanct, so it’s not clear they can be deposed and it is very clear they cannot be assaulted or dragged. Violating the sacrosanctity of a tribune is, at least notionally, a capital offense and a severe violation of religion and if you think that Tiberius Gracchus’ legal basis for all of this is rubbish, you think he just did it twice. Of course, Tiberius is also a tribune, so you can’t attack him now, but once his year is done, you are probably planning to haul him in to court and let a jury decide if what he did was legal or not.2
In any case, with Octavius removed, Tiberius passes his land reform bill. The law provided for a three-man commission to handle the assessment of what public land was held in excess and then to hand it out. Tiberius Gracchus names as those commissioners himself, his brother and his father-in-law (Appius Claudius Pulcher (cos. 143)). Needless to say, that is a set of commissioners which does not inspire a lot of confidence that the commission will be uncorrupted by politics, a point we’ll get back to in just a moment.
In the meantime, the Senate looked to exert its traditional prerogative over state funds (as it advised the quaestors who superintended the treasury) to hamstring the new commission, but Tiberius Gracchus took advantage of the recent death of Attalus III, King of Pergamum. Attalus had notionally willed his kingdom “to the Roman people” – he had no clear heirs and so perhaps thought by this act to get the Romans to pick one of his relatives to run the kingdom, thus avoiding a damaging civil war – but instead Tiberius, getting the news early, rushed to pass a law annexing the kingdom and using the windfall to fund his commission. The law passes, but this is a breach both of the Senate’s traditional power over state finances, but also its very important role managing Roman foreign policy.
What I want to note in this sequence which is important for understanding what comes next is that Tiberius Gracchus has just demonstrated that, so long as he remained popular, he could use the powers of the tribunate to essentially run the Roman state from the tribune’s chair. Tiberius has now forced not merely a domestic land issue, but also a finance issue and a foreign policy issue over the objection of the Senate and another elected tribune, essentially running roughshod over all of the customary limits intended to keep any one Roman politician from coming to dominate the Roman political system.
Of course if you were an opponent of Tiberius Gracchus, you could at least tell yourself that this is all bad, but at the very least, Tiberius Gracchus will be out of office next year, as it was contrary to custom to run for any office immediately after holding it. Indeed, it was unusual to hold basically any office more than once, save for the consulship (and even then, only for very successful consuls and never multiple years in a row). Those limits are customary but everything about the Roman Republic is customary; if you discounted the mos maiorum, there wouldn’t be any republic left. You’d instead expect that Tiberius would go back to being a senator for a few years while planning his shot at the praetorship – during which he’ll have to survive a series of court battles over the legality of his actions.
So even if he is doing potentially outrageous, dangerous things, at least he’ll be gone in a year, right?
Bret Devereaux, “Collections: On the Gracchi, Part I: Tiberius Gracchus”, A Collection of Unmitigated Pedantry, 2025-01-17.
- Which, again, noting the complications above, probably means applying that limit for the first time to at least some classifications of land it had not applied to before and also applying it against the socii.
- The Roman court system leaves questions of law – which in most modern courts would be decided by a judge – to the jury itself.
June 4, 2026
QotD: Demographic decline in the late western Roman Empire
It’s a plausible theory and the only problem with it is that it doesn’t appear to have actually happened.
Here the key archaeological method is what is called “field survey“. While readers are probably more familiar with the intensive excavation work done at famous sites like Pompeii or Vindolanda, one tool archaeologists have to study the past is to survey large areas, sometimes by air, sometimes by on foot, sometimes with ground penetrating radar, in an effort to map out larger scale settlement patterns in the past than would be possible by labor-intensive single-site excavation work. Dateable remains (pottery most often) allow for archaeologists to get a rough sense of the dates in which sites were inhabited and in some cases building remains and the like can give some sense of what kind of settlement was present. The “error-bars” on some of this data can of course be large, but they offer a tool for tracking long-term changes in land use patterns. On the flip side, these sorts of studies really become valuable only when you have a lot of them to create a robust data-set over a fairly large area that lets you adjust for purely local patterns and distortions. Fortunately in much of the former Western Roman Empire and especially in Roman Italy (where these studies are very important for the study of Roman demography and agriculture) we’ve hit the tipping point where there is enough archaeological data to begin reaching for conclusions.
Now there is an immediate difficulty with using this kind of evidence, which is that for reasons we’ll get to in a moment (though they are reasons that tend to also be bad for the “change and continuity” argument), we have a major confounding variable here: site visibility. Our ability to see a site, archaeologically, is heavily dependent on factors like building material and the quantity of imperishable goods (especially pottery) that people are using. For reasons we’ll get to, compared to, say, second century AD communities, sixth century AD communities tended to build their buildings in far more perishable (and thus less visible) materials (like wood) and also tended to use a lot less imperishable household goods. Consequently, it is substantially harder to see a sixth century village than it is to see a second century villa.
Nevertheless, the decline is so marked and so consistent as to strongly suggest there is something real here. R.P. Duncan Jones (in “Economic Change and the Transition to Late Antiquity” in Swain and Edwards (eds) Approaching Late Antiquity (2006)) assembles some of the site data from around the empire; there is unsurprisingly a lot of regional variation (with some regions, like Syria, actually moving against trend), but in the western Empire (except N. Africa; decline there comes later) the trend is fairly clear, with site numbers declining (often drastically by half or more) beginning in the late third or fourth centuries. Bryan Ward-Perkins in The Fall of Rome notes a field study outside of Rome in which the number of sites declines by three quarters. Site data accumulated like this isn’t often very chronologically precise, so we’re dealing with centuries, not decades, but the clear trend suggests rural population decline, not an urban population ruralizing. To be visible to us in this way, the decline must have been quite severe.
To give a sense of the scale of the decline, here is an abbreviated version of a chart from Bruce Friar’s “Demography” chapter in the second edition of the Cambridge Ancient History, which breaks down the estimated population of the Roman Empire by region and adds the dates when each of those regions got back to their Roman-era population:
Chart from Frier, “Demography” CAH2 XI (2000), 814. Some of these figures would likely see some revision today, mostly downward revisions of growth combined with upward revisions in population reflecting a somewhat (but generally not massively) higher estimated pre-Roman population.
Note that the decline in the East was, as noted last time, both later and generally slower. The reason for the later times to reattain Roman population here in many cases is that the major medieval Islamic population centers were further East (e.g. Baghdad under the Abbasids) placing them outside the traditional bounds of the Roman Empire, but also that the Roman East was much more urbanized and densely populated compared to its land area than the Roman West in the second century (or at any time during the Roman Period) so the “population to attain” bar on the East was much higher. After all, the cities of places like Syria or Egypt were in many cases centuries or even millennia old when the Romans showed up.Now the long times there to regain the Roman population can be a bit deceiving (and are very approximate). For reasons we’ll get into shortly, population growth from 600 to 900 or so in Europe was very low, so the issue here isn’t that the decline was so steep that it took many centuries to recover from, but rather that the decline was from a high population equilibrium to a low population equilibrium, both of which were, under their own conditions, stable (if that is confusing, don’t worry, we’ll delve more into it in a moment). Second, the apparent gap between places that “caught up” before 1300 and those that “caught up” after it is smaller than it looks, because of course the mid-1300s represent a massive population discontinuity over the entire broader Mediterranean world due to the Black Death such that a lot of those places “catching up” in the 1200s probably fell behind again due to the plague and then caught up again in the 1400s or early 1500s.
But this now raises two related questions: first, why did population decline so sharply and second, what was the impact on quality of life that resulted? The old answer to the first question was of course “the barbarians killed everyone” but as we’ve seen, while the fifth century was a violent time, the violent discontinuities were not that extreme. Surely the violence of the period has something to do with some of this declining population, but as noted, the underlying population (with their language and religion) didn’t much change (and the raw number of “barbarians” coming over the frontier was, in demographic terms, fairly small). Most of those Roman cities decayed, rather than being burned. But if the “barbarians” didn’t kill everyone, what did and why did that somehow have a negative impact on the survivors? The answers to these two questions are actually linked in that they depend on the same evidence, so that is where we will go next.
Bret Devereaux, “Collections: Rome: Decline and Fall? Part III: Things”, A Collection of Unmitigated Pedantry, 2022-02-11.
