Welcome everyone to the third every Coliseum debate. This one is Thanas V Duchess of Zeon on Luttwak's theory on the Grand Defensive Strategy of the Roman Empire. Thanas is arguing against Mr Luttwak's work while the Duchess of Zeon will be defending it. As before first the original two challenge and acceptance posts for the record.





Second as part of the standard procedure for the forum the following rules will be observed during this match as agreed between the two parties.



In short this is a first for us in this forum as we will not be debating an idea, or a belief system but rather a third parties work. I should note at this point that as the moderator of the forum I have made one change to the initial OP as provided to me by Thanas that being the offer of disscussion of the thread in a provided History Comment thread. Such a thread will be created but only after the debate has come to an end on January 7th. I should not as well we will be going with a response, counter-response format with the last posts being accepted on January 7th at 2359 Eastern Standard Time. At that time a comment thread will be created but until such time any mid-match comment thread will be deleted.

And without any further delay lets get things rolling.

I would like to thank Mr. Wong and the Mod team for allowing me to have this chance, and my honorable opponent for suggesting the topic of this debate.


First, let me point out that Luttwak deserves a lot of credit for even attempting such a study and thereby forcing Historians to re-examine the subject. In the words of Benjamin Isaac: “If we do ask this question, it is thanks to Luttwak's own admirable synthesis. However,[...] [his hypothesis] is not based on an independent analysis of the ancient literary sources or the archaeological material, but derives from a lucid perusal of modern literature[...]it cannot be proved or disproved by objective means. What we can do, however, is to test the basic assumptions which underlay the hypothesis by referring to ancient sources and archaeological material.” [1]

As will be shown throughout this post, Luttwak's study suffers from several mistakes which sadly invalidate the enormous worth of his work.


General Problems with Luttwak's study

Luttwak's study starts at a time when the Roman empire was already formed, namely 23 BC and ends with Diocletian, well before the end of the western empire. In doing so, it ignores the long period of Roman expansion under the Republic, as well as the later changes made to the system by Stilicho and other late Roman generals. In fact, of the fifteen centuries in which a Roman empire existed, Luttwak has decided to merely examine four, while ignoring the “strategy” of the Roman Republic.[2] As will be shown in this text, Luttwak therefore misses important clues that seriously undermine his thesis, but these will be handled when discussing the three systems.

The question is whether the Romans had any concept of Grand strategy at all. This is questionable. Roman emperors allocated legions in empire-wide movements and therefore thought at least in empire-wide strategic terms, as Kimberley Kagan has shown [3]. However, the question remains whether the Romans had a generation-spanning Grand strategy to defend or expand the empire. Certainly, there was no ancient equivalent to West Point in ancient Rome. We find several institutions that might have formed an institutional memory, for example the tribunes vacantes, officers without assignments attached to the emperor or field commander.[4] Other such institutions might have been the imperial bureaucracy in charge of rationing or weapon production.

Yet this does not help to settle the question whether the Romans had a concept of Grand strategy, or even of strategy. As Benjamin Isaac shows, the word strategy is a modern invention. For example, while Thucydides gives various reasons for the move of Brasidias into Thrace, it was Liddel Hart who first described it as a “strategy of indirect approach.”. This form of abstraction and rationalization is a modern concept not found in ancient sources.[5] Against this idea, Everett Wheeler has argued that the Romans used the alleged synonym consilium instead[6], yet of all the examples he uses only one may serve as any indicator of grand strategy - the advice to Tiberius to stop further expansion “consilium coercendi intra terminos imperii” (Tac. Ann. 1.11.8), but I doubt anyone would consider it to be grand strategy instead of the warnings of an old emperor about to die. Tiberius clearly disregarded that advice when he ordered Germanicus to invade Germania. Furthermore, all Romans thought expansion to be the natural course of the empire.[7]

In any case, the debate of the existence of Roman strategy is an exhaustive one, and to go into great detail would shift the focus of this debate too much. This author leans more in the direction of Isaasc, a noted specialist on ancient frontiers, than in the direction of Wheeler. However, the author admits that he has not come to any definite conclusions on the subject and is therefore willing to stipulate for the purposes of this discussion that whether accidental or by plan, the Romans at least acted in a strategic manner.

Luttwak argues that the deployment of forces primarily had the focus of defending the empire. This is more than questionable. For example, J.C.Mann argues convincingly that the primary motive of the emperors was expansion[8]. Furthermore, Luttwak ignores the integration aspect of the army. Auxillaries, for example, served the purpose of integrating the subjugated peoples as well as providing the empire with specialist forces or more manpower. In the east, the Roman army in Judea primarily served as a peacekeeping force, as Isaac has shown – Roman troop deployment in Judea was dictated by the necessity in face of revolts or political developments, not by an outside threat.[9] The Roman army assisted with tax collection, infrastructure projects and performed administrative functions.[10] Even if the main function of the Roman legions was to fight, in reality they were essential to the Roman provinces as well, for example Roman military architects built aqueducts, soldiers guarded prisons etc.[11], something Luttwak ignores in favor of concentrating on the military aspect only.

The question is now whether Luttwak's systems accurately describe at least the status of the Roman defense. And therein lies the next problem – oversimplification. As the foreword by Prof. Gilliam in Luttwak's book notes, specialists will “doubtless find new errors” and historians will “react to abstract patterns by pointing out complexities”. However, Luttwak makes no real effort to evade such criticism – in fact, the depiction of his models between the systems (the first showing an empire surrounded by numerous clients, the second showing an empire with no clients at all), is not only a dangerous oversimplification of history, but a distortion of Luttwak's own words as well.[12] It would also have been only proper if he had given a source for his reconstructions of border fortifications instead of simply drawing diagrams.

In his preface, Luttwak further makes a number of claims that a historian must find issue with. There is the unproven assertion that Roman tactics were not superior to those of the enemy, the assumption that the imperial Roman soldier was not noted for his élan and finally that Roman weapons were not regularly superior to those of the enemy. All those claims are questionable at best, and since Luttwak does not provide any evidence for them, it would be best to discount them. Another general criticism is that Luttwak does not rely on primary sources, but that he reinterprets the findings of other historians indeed. There is not a single new source which is revealed in his book.

Having said that, it is time to go in media res, to examine the various systems.


The first system: Client defense



This system primarily rests upon the assumption, that indirect rule was favored to direct rule in that period, and the majority of Luttwak's examples deal with the east of the Roman empire. It is therefore extremely devastating to his theory that a new work [13] on the Roman east does not support any of his conclusions. Kimberley Kagan [14] has already aptly summarized the points made in this study, which deals primarily with political and social factors, but which are devastating to Luttwak's underlying assumptions. I will now expand on her summary whenever the east will be discussed:

1. There was no single trend towards direct or indirect rule. Both Augustus and Tiberius alternated between the two forms according to the situation at hand.[15]
2. The Roman military presence increased gradually and in response to different occurrences[16]

A closer examination of Millar's work reveals another problem with Luttwak's theory. There is no direct evidence to assume that the client states were profitable.[17] In fact, when cappadocia was provincialised, the Romans expected a huge raise in revenue[18], something that had not been available as a client. The final argument against the efficiency of the client system is the fact that the Romans provincialised almost every client kingdom.

Although Luttwak posits an empire-wide defense system, which would be inapplicable now, the author believes it would be premature to conclude that his whole theory is wrong. What about the western part of the empire? Luttwak posits here that the Romans used clients although they were unreliable. But did the Romans use clients in order to bolster the defense, clients who turned a net profit for the empire?

In fact, this is not the case. The clients on the northern frontier caused considerable trouble for Rome. This is highlighted by the first revolt (9 AD) which destroyed three legions and caused the empire to commit to a lengthy offensive campaign of eight legions, which was abandoned by Tiberius. The rebellion of the Batavi (69) caused the loss of four legions to varying causes and lead to widespread devestation. The Frisii, Rome's most dependable client in Germany, revolted 28 and 47 AD, causing losses that were so high the Emperors did not dare to publish them in Rome.

Translation and text taken from here.

Furthermore, it is more than questionable if the Romans ever pursued a client policy in Germany as they did in the east. The practice of resettling allied or subjugated tribes (Ubians, Treveri) or placing legions directly in their territory is anathema to Luttwak's theory. Therefore, if the Romans used a client system in the west at all, it is more than debatable if it worked according to Luttwak, by reducing the strain on the empire. In fact, legions always had to be stationed near the clients.

Based on this argument, it is far more likely that the Romans used clients as they saw fit. To say that clients were a net gain to security and resulted in vast disposable military strength is a falsehood, as the management of clients and the security of the clients themselves consumed manpower. Luttwak's theory may apply to some clients, but certainly not to the vast majority of them.


The second system: Forward defense



Having previously shown that the clients cannot be really seen as being the prime element in the Roman defense system, it now is the time to decide what can remain from the second system. Of course, the notion that the abolishment of the clients generated a huge stress on the empire's resources must be dismissed.

Likewise, the notion that this system had no disposable surplus strength has to be dismissed. In fact, during the time period, 11 legions were formed. (II Adiutrix, IV Flavia Felix, XVI Flavia Firmy, – Vespasian; I Minervia – Domitian, , II Traiana, XXX Ulpia Victrix – Trajan; II, III Italica – M.Aurelius; I, II, III Parthica – S. Severus). It is also questionable if the Roman empire only had a brief “surge capacity”, as Luttwak asserts. This is easily countered by the fact that the new legions were permanent formations. Furthermore, a comparison of Hadrian's actions and the actions of Augustus after the loss of the Varian legions is not as unfavorable as it might seem at first. Unlike Augustus, actual core territory of the empire was threatened by the Bar-Kochba revolt (Egypt and Syria), and unlike Augustus, the rebels were defeated and the lost territory recovered. Finally, Hadrian did expand into Arabia and did not halt the expansion of the empire.

The author agrees with Luttwak that defence is forward in nature. Yet this is nothing new – the limites (fortified roads into Barbarian territory) existed since the time of the first invasion into Germany in the west. Therefore, it would be wrong to conclude that forward defense is only introduced in this system. Also, forward defence was not the primary aim of the Roman army in every region. As has been shown by J.C. Mann, the aim of the empire was expansionist in nature[19]. Marcus Aurelius reportedly wanted to annex Marcomannia and Sarmatia as provinces. Usually, new legions were created when new provinces were created, instead of trying to bolstering the frontier, as Luttwak claims.[20] Finally, it is in this time period that we see the largest extent of Imperial power. There is no evidence that the Danube frontier was overstressed, as Luttwak claims[21]. Instead, the recruitment of new legions was usually tied to the establishment of new provinces (see 20), and in this case due to the need of new legions after the old ones had suffered from the plague. It is not an indicator for the general situation on the Danube at all, but rather an indication of the enormous strength the roman empire showed when confronted by an empire-wide plague.

It would be wrong to assumes that this system was not any more flexible than the “previous” one, for so far they both seem to have the same structures. In fact, the system proved its elasticity many times. Luttwak makes the mistake to assume that just because legions did not change their base frequently, they themselves must have been rather static in nature. In fact, Roman legions had become even more flexible in unit organization. The widespread use of the formation known as the vexillatio now permitted individual detachments from one legion to become attached to other legions. Allow me to demonstrate this flexibility with regards to the persian war of 161.[22] In the autumn of 161 AD Parthian troops attacked Armenia, conquered it and defeated the governor of cappadocia. Two legions, the only ones stationed in Asia Minor, were defeated. However, the Roman response was swift. Syria, which had three legions, detached troops to defend Asia Minor, whereas Lucius Verus arrived with vexillations from the Danube. Through these vexillations, the strength of the legions was increased to the equivalent of six legions in total. By 162, the Parthian offensive had been stopped, and eventually the Romans conquered the parthian capital itself in 165, a feat not achieved under the first system, but which occured thrice under the new system. Following the empire's success in dealing with the deadliest enemies of the period, it might be argued that in this timeframe, Rome was even stronger than under the Julio-Claudians.
It is also worth noting that the Legion numbers in a province were already largely static since Claudius, and that whole legions moved even during the second and third century AD, further contradicting Luttwak's theory. [23]

Having said that, the most important question arises – did the Romans have a concept of a static or a firmly drawn frontier? Luttwak seems to believe so. Yet Isaac has shown via an exhaustive comparison of the literary evidence that Limes does not mean a fixed line or a series of fortifications .[24] Archeological evidence shows that on many “frontiers”, the Romans did not have any garrisons at the frontier at all. In the case of Tripolitania, the earliest garrison may be found in the time of Commodus. And in the case of Mauretania Tingitana, which was garrisoned by auxillaries, no frontier line exists at all. Likewise, in Dacia, the frontier did not consist of heavily-garrisoned forts on the border, but of detachments spread throughout the province, most likely for supply reasons.[25]

It is also more than questionable if the Romans were in a search for natural/fixed borders or even defensive in nature in this time frame. As Kimberley Kagan argues, the Rhine, Danube and the Euphrates could denote natural borders – yet imperial fortifications ran to the east of both Rivers, Trajan conquered Dacia and Mesopotamia was several times annexed and lost during the course of Roman history.[26] It would also be wrong to present those cases as defensive wars, as Luttwak does with Dacia. The burden of proof is on the side who argues in favour of the defensive measurements, for we know that the Romans had a long history of expansion. Propaganda pieces and literary evidence show that the military glory of the emperor and the expansion of the empire was the most important reason for war. Or, in the words of J.C.Mann[27]:

Another argument against fixed lines is the fact that the Roman border in Britain moved back and forth several times. Severus campaign in Britain is a case were Cassius Dio claims that it was offensive in nature, echoing Sueton's claim of Domitian's assault against the Chatti.[28]
Finally, the most damaging aspect against Luttwak's second system once more arises from the east. Using extensive epigraphical and literary evidence, Millar shows that from the reign of Vespasian, the Imperial Army in the east began an extensive building program and expanded eastwards and southwards. When reaching natural borders (the Euphrates and the Taurus mountains), the Empire began wars of conquest. Trajan invaded Armenia and Mesopotamia. Hadrian returned Mesopotamia to the Parthians since he had to deal with the Bar Kochba revolt in Judea. After the revolt, he annexed Arabia and engaged in several administrative works and road building in Judea, Syria and Arabia. In the war of 161-165 the Parthians lost Dura Europos, as well as Osrhoene, Nisibis and Singara (of course, Septimius Severus' campaign against Parthia resulted in the complete destruction of that state as well as the establishment of numerous garrisons in formerly “parthian” territory). In the same time frame, in which Luttwak claims a defensive empire and a period of forward defense, the Roman empire expanded aggressively in the east and consolidated internal control of that territory.[29]

Therefore, the second system has to be largely dismissed as well, with the only salvageable part of it being the idea of forward defence.


The third system: Defense-in-depth

Considering that the first two systems are lacking, it might only be appropriate to not even discuss the third system. Yet this is the most interesting of all three – the idea that the Roman empire developed a defense-in-depth system.



First of all, it should be pointed out that the posture of the empire, as explained above, was not primarily defensive in nature and there is no evidence that the ideology of the empire has changed in that regard. Nevertheless, throughout the third and fourth centuries, the empire is heavily involved in defensive actions. It is easy why one should therefore conclude that the Romans invented a strong defence system. They did so, however it should be noted that this is not necessarily one of defence-in-depth, but as the author will show, an evolution of the concept of forward defense.

It is quite unlikely that the Romans had the geographical skills necessary to formulate a defense-in-depth. Roman maps are not made according to topographic reasons, but according to travel distances between cities, nor did the Romans think in concepts of territories, but rather in concepts of population in their decision making.[30]

There is no clearcut archeological evidence to support a change from forward defense to defense-in-depth. Luttwak claims that Diocletian supported a “shallow form of defence-in-depth” - which is a contradiction – yet J.C.Mann shows that this did not result in a defence-in-depth. As he says[31]:

Furthermore, he shows, that all legions stationed in Judea and in Syria have been moved up to the frontier. If anything, the chances of a defence-in-depth are decreased by such a move. Likewise, considering the roads in Britain, it is clear the primary function of the units stationed in the Hinterland was a forward defense or an attack made through the frontier, not a defensive move against marauding barbarians.[32] Modern maps or studies on the subject show Roman forts build directly on the boarder, not in the hinterland. Furthermore, Roman forts were still positioned in enemy territory, on the other side of the frontier.[33] This seems once more indicative of forward defense rather than defense-in-depth.

What concentration of troops would have stopped a penetrating army? And this is where the cut-off point chosen by Luttwak for his study is shown to be too early. The only armies that could have stopped invading barbarians quickly enough were the regional field armies, most of which were not formed until well after Luttwak's period of choice (the field army of Illyrium, for example, might have been formed as late as 409 AD). Stilicho's practice of creating small reserve armies is also way too late for Luttwak to notice it. Yet if we assume that those had the objective of stopping barbarians instead of internal security (something that seems to be the case when regarding the preasentalis I,II), their deployment in relation to the limitanei would be very similar to those of the old legions in relation to the limitanei.[34] In fact, this is another indication of forward defence – a heavy increase in border troops, quality-wise as well as quantitative.

Luttwak tries to argue that the forts in the province of Paleastina III may have served in such a function, but as Isaac has shown, their existence is just as easily (and more convincingly IMO) explained by the need for internal security.[35]
Millar shows that during the time of the third century, no system of defence-in-depth was in place or was invented against the persian incursions. In fact, as Millar shows, it was only after the crisis of the third century that the strata diocletiana denoted a formal frontier, yet its purposes were to achieve internal security as well as external. No system of defence-in-depth therefore can be found in the east.[36]

In any way, what little archeological evidence exists, does not support that there was defence-in-depth alone. There is a definite need for literary sources to back that theory.

The literary sources, however, do not mention defence-in-depth. In fact, they all proclaim forward defence. In the third century, a host of ancient writers, including Herodian, mention extensive offensive campaigns against enemies. One such battlefield of the third century was recently discovered in Germany, in Kalenfeld-Oldenrode, Link to german video, 200 miles to the east of the Saltus Teutoburgienses, where Varus perished. That the Romans conducted such offensive operations with at least one region in the third century seems to be a strong indicator that the claims of Ammianus and Herodian with regards to the extent of Roman operations should be taken as fact. Furthermore, it shows that Roman army units operated to the east of the Rhine in army strenght, and is therefore another factor against the defence-in-depth theory. If the theory I have read is correct and the Romans tried to intercept plunderers from the area of Jutland[37], this would be another argument against defence-in-depth.

Another factor against the theory of defence-in-depth is the fact that in the third century, marauding bandits could mostly only be stopped after they tried to cross the border back into their lands. This is exemplified by this inscription (AE 1993, 1231, found in Augusta Vindelicorum/Augsburg):

As you can see, this inscription from the year 260 AD talks about how soldiers of the province raetia liberated many thousand Italians from captivity of marauding barbarians. The fact that the Barbarian raiders were able to raid or even reach Italy and then return back, laden with booty and prisoners on foot, speaks against a system of defence-in-depth.

But since Luttwak claims that Diocletian finalised this system, maybe one should only regard evidence of the fourth century as sufficent. Yet the campaigns Ammianus Marcellinus writes about are distinctly similar to forward defence. Let us review what Ammianus writes about a succesful Roman operation against the Sarmatians (Amm. Marc. XVII, 12):


Translation taken from here.
As one can see, this should have been exactly the kind of attack a defence-in-depth should have prevented – small, plundering detachments. Yet there is no mention of such a defence. In fact, what decides the issue are two fast attacks on the barbarian hinterland. The decisive battles all take place outside the empire. Unlike other writers, Ammianus was also a staff officer, so his account and his omission of any defence-in-depth has to be taken seriously.

Likewise, almost every other campaign by Ammianus Marcellinus has the hallmarks of forward defence, whereas none that is known to the author speaks of a defence-in-depth. The author is willing to give more examples (for example, an analysis of the defence of gaul under Julian, or of the gothic invasion of Adrianople) and if such a request is made, it will be followed as soon as possible. However, for the sake of brevity (the analysis is over 35 word-pages long), the author feels that he has taxed your patience enough with the above-mentioned examples.

The author therefore feels that he has succesfully argued that Luttwak's theory cannot stand without substantive modification, if it can stand at all.


References and notes:

[1]Isaac, B., The Limits of Empire. The Roman Army in the East, 2nd rev. Ed., Oxford et al 1992, p. 377.
[2]A polemic might therefore conlude that the better title for his work would have been “The Grand Strategy of a part of the Roman empire.
[3]Kagan, K., Redefining Roman Grand Strategy, The Journal of Military History (JMH) Vol. 70 (2006), pp. 354-361. Kagan largely draws for her analysis on the excellent article by Ritterling, E., “Legio”, in: Realencyclopädie für Antike und Christentum 12, Stuttgart 1924-25, col. 1186-1837.
[4]For the tribunes vacantes see Ammianus Marcellinus: „In hac multiplici virorum inlustrium clade Traiani mors eminuit et Sebastiani, cum quibus triginta quinque oppetivere tribuni vacantes et numerorum rectores et Valerianus atque Aequitius[...]“ - Ammianus Marcellinus, 13.18 (“Amid this manifold loss of distinguished men, the deaths of Trajanus and Sebastianus stood out. With them fell thirty-five tribunes, without special assignments , and leaders of bodies of troops as well as Valerianus and Aequitius[...]” Translation by the Loeb edition.
[5]Isaac, Limits, p. 375.
[6]Wheeler, E., Methadological Limits and the Mirage of Roman Strategy Part II, in: JMH 57 (1993), p.217
[7] Mann, J.C., Power, Force and the Frontiers of the Empire, Journal of Roman Studies (JRS) Vol. 69 (1979), pg 176-177
[8] Ibid.
[9] Isaac, Limits, p. 54-160.
[10]Ibid, p.101-160.
[11]For a good summary see Le Bohec, Y., L'armée romaine. Sous le Haut-Empire, Paris 1990, ch. 3.II,III as well as the introduction to the book.
[12]Ibid, Luttwak p. 23,24.
[13]Millar, F., The Roman Near East 31 B.C. - A.D. 337, Cambridge (Mass.) 1993.
[14]Kagan, pg. 344-5.
[15]Millar, pp. 38-56.
[16]Ibid, pp. 56-70.
[17]Ibid, pg. 49-52.
[18]Tacitus, Ann. II., 56,4, Millar pg. 52.
[19]Mann, Power, pg. 177
[20]Mann, J.C., The Raising of New Legions during the Principate, in: Hermes. Zeitschrift für klassische Philologie Vol 91 (1963), 483-489.
[21]Luttwak, pg. 126.
[22]The details follow Cassius Dio LXXI 2,1-3; as well as Syme, R., Avidius Cassius. His Rank, Age and Quality, in: Historia-Augusta-Colloquium Bonn 1984/5, Bonn 1987, pp. 207-222.
[23] Ritterling, E., “Legio”, in: Realencyclopädie für Antike und Christentum 12, Stuttgart 1924-25, col. 1186-1837
[24]Isaac, B., The Meaning of the Terms Limes and Limitanei, in: JRS 78 (1988), pp. 125-147.
[25] Mann, J.C., Power, Force and the Frontiers of the Empire, Journal of Roman Studies (JRS) Vol. 69 (1979), p. 179-180.
[26] Kagan, p. 342.
[27]Mann, J.C., Power, Force and the Frontiers of the Empire, Journal of Roman Studies (JRS) Vol. 69 (1979), p. 178.
[28]Isaac, B., The Limits of Empire. The Roman Army in the East, 2nd rev. Ed., Oxford et al 1992, p. 387-390.
[29]Millar, pp. 70-141; Kagan pp. 344,345.
[30]Isaac, pg. 401-408.
[31]Mann, Power, pp. 180,181.
[32]Ibid, p.181.
[33]Cf Elton, H., Warfare in Roman Europe. AD 350-425, pp. 157-161 as well as the huge study by von Petrikovits, H., Fortifications in the North-Western Roman Empire from the Third to the Fifth Centuries A.D., in: JRS Vol. 61 (1971), pp. 178-218.
[34]Mann, p. 182.
[35]Luttwak p. 160, Isaac Frontiers pp. 162-171.
[36]Millar, pp. 159-73.

Due to unforseen circumstances, the Duchess of Zeon has requested this match be placed on indefinite hold. Assuming Thanas agrees this debate topic will be held in reserve until both parties are ready to resume.