In recent decades, a number of remarkable and counter-intuitive claims have been advanced about the Roman and early Byzantine imperial borders. We have been told that the empire had no clear or fixed borders or even no concept of a border to begin with; that there was no expectation that the borders could or should be defended and no actual ability to do so; that there was no imperial strategy for the defence of the empire; that there was no conception of the territorial integrity of the Roman state; that the border was always permeable, porous, and fluid; and that features of the terrain were not used as borders or imagined as marking the border. Not all of these theses have been advanced in the same publications, but as a coherent constellation they have given rise to a revolutionary understanding of the imperial borders that is often encapsulated in the catchphrase ‘fluid zones of interaction’.

Geoffrey Greatrex has effectively refuted most of the claims made above, with special reference to the borders of the late eastern Roman state (also known as early Byzantium), and many other historians have signalled their doubt regarding this picture or key components of it. In this chapter, I will first present some of the conceptual weaknesses of the idea that borders were ‘fluid zones of interaction’, a phrase that is now routinely applied by many scholars to Byzantine borders of all periods. In the second part, I will focus on the borders of the middle Byzantine period and argue that, even if they were sometimes porous and fluid, that was not what defined them as borders. All places are porous and fluid absent a force of constraint, therefore borders are better defined as sites of potential intervention by the state to preclude movement. Borders were created, maintained, and regulated by specific state institutions, and any fluidity or interaction that took place across them was usually by their sufferance, indifference, or failure, and not because they did not exist or were completely ineffectual. Premodern state institutions were not as impotent as many think. This means that borders must be defined primarily in terms of institutional practices, and their porousness evaluated against the success or failure of policy goals.