Land and Military in the Roman Empire
It is often assumed that thematic system of military lands was a wholly novel measure, and that later pronoia grants were significantly different from previous stratioka ktemata. Yet neither assumption is likely to be true, even if there is no direct evolutionary line leading from prata legionis to stratioka ktemata and pronoia. This will only cover the first part of the question.
Prata Legionis / Territorium Legionis
Legions during Roman Empire typically used local provision for supply. Aside from buying food etc. from local producers, they also grew some of their own food. This was possibly grown on land given to the army. Roman auxilliary forts and legionary fortresses were surrounded by clearly demarcated land which was termed either prata or territorium. Territorium might have referred to all territory administered by the legion, while prata was likely only referring to pasture (grazing land for animals). In Spain, inscribed boundary stones marked the division between prata legionis of Legio IV and the lands of Juliobriga and Segisama. Similar boundary stones existed at Burnum. This area could be significant: Legio III Macedonica had over 600 square kilometers in its prata legionis. For a legion of 6 000 soldiers, just producing wheat would require 3 500 ha of land.
Prata legionis were mostly pastures and were used for horses, draught animals and animals for food – potentially over 5 000 individuals. Feeding this number of animals would have required several thousand hectares of land, and, with auxilliary forces, several tens of square kilometers just of productive land. Prata legionis served as arable land, pasture and other economic and logistics functions, but also for army maneuvers and training.
Aside from land given to legions themselves, there existed a practice of granting land to settlers on condition of their serving in the legions. Smaller units – alae, independent cohorts – also had their own land.
Lastly, legionary fortresses themselves were often surrounded by canabae, civilian settlements surrounding legionary camps formed by traders who sold supplies by legions. These settlements also helped facilitate contact between legions and local populace. Eventually, retired veterans started settling there with their families. Canabae had status separate from civilian settlements and were governed according to their own laws.
As early as Principate, legionaries started acquiring land. As legions spent long time in one place, not only did legionaries started acquiring land, but legion also became immersed in local society, often providing administrative duties. This immobility of army units likely led to Diocletian establishing Comitatenses and Exercitus Praesentalis, as provincial legions could not be moved without causing massive disruption to local society.
Borders were typically situated on rivers – borders towards barbaricum and between provinces alike. For this reason, as well as for reasons of resupply, logistics and watering, prata legionis was typically near river.
Landowining soldiers are well known in the Exarchate of Italy from the sixth into eighth century, and indeed exarchates may have been based around landed soldiers and thus a model for later themes. There is, however, little evidence for this.
Stratioka Ktemata appears to have evolved gradually. In its final form, it was land owned by soldiers who drew much of their finance from said land and lived on it. But originally there appears to be little difference between Late Roman and Middle Byzantine logistical arrangements. In fact, logistical arrangements appear to be based on coemptio, compulsory purchase of military supplies, often done in kind.
What appears to have happened (as described in The Empire That Would Not Die) is that armies were first withdrawn to a defensible frontier in Anatolia. Each army was assigned territory according to its productive potential and fiscal capability in proportion to number of soldiers in army. Said territories were merely groupings of Late Roman provinces. By late seventh century, these groups of provinces came to be known collectively by name of the army based there. In order to supply armies, fiscal arrangements were changed. Instead of raising tax in cash, which was then used to purchase supplies, tax was raised in kind, with resources going directly to the army. This system mirrors that already established by Diocletian. To facilitate this, soldiers had to be distributed throughout the provinces, instead of being stationed in few central locations (camps etc.).
System as implemented appears to have been based on hospitalitas system as employed in housing foederati. During settlement of foederati, barbarians received land fractions in an extension of Roman billeting law. This did not involve ownership of the land; rather, soldiers were funded based on designated unit of assessment.
Soldiers continued to be paid by the government. They did eventually come to provide their own equipment and depend on their own property, and soldier is even early on understood to have been supported by the household. It was a natural consequence of reduced salaries and less frequent payments, much of which had been come to be delivered in kind instead of in coin. Already by 8th century it was assumed that soldiers own their own weapons, armour and horse. Armies continued to rely on established pattern of support, by which provincials delivered certain quantities of grain and other produce. Soldiers do seem to have been granted land from imperial estates, which disappear from records after 7th century, indicating that already described was in fact a centrally planned and not organic process.
At any rate, extreme decentralization and localization of government and military was one of its main strengths. Tax system was highly informal and flexible, with very differing implementation at local level. This localization, and rooting of system into local society (including localized recruitment), allowed it to adapt to changing circumstances and localized as well as widespread crises.
As can be seen from the above, practice of localized military support has long tradition in Roman military, and it is precisely in this tradition that origins of thematic systems should be sought. Rather than creation of a wholly new system (as proposed by Ostrogorsky) which in conditions of 7th century was likely impossible to achieve, thematic system of miltary lands presents a gradual evolutionary development of military organization that has its roots in military arrangements of early Principate.
The process which happened with Principate forces likely also repeated itself after 7th century. Originally merely field armies billeted across Empire for sake of simplifying logistical support, themes eventually came to merge with civilian society and take over much of duties of provincial administration. Soldiers themselves also came to acquire land (a much safer source of income than coinage), and thus became part of local society. Evidence does exist that this was planned, and not natural process, however.
Whatever the case, acquisition of land did lead to increased immobility of thematic forces. While heavy cavalrymen – kataphraktoi – and possibly cavalrymen in general had estates too large for one person to handle, and thus their absence was not crucial, light infantrymen could not be so cavalier about campaigning. It is likely in this that we have to seek causes of increasing favorization of imperial Tagmata as well as Nikephoros II’s preference for heavy cavalry, which could campaign outside borders of the Empire – though when it comes to favoritization of Tagma units, countering rebellions by themes was also a significant concern.