Commentators from ancient Rome to the fifteenth century liked to wax poetic on the importance of western Europe's waterways for short- and long-distance travel. But as Marjorie Nice Boyer--whose work on medieval transport is still being republished more than fifty years after the fact--pointed out, a lot of those writers had never actually been to the places they claimed to describe. It's true that the question of ongoing use of the Roman road system from the early Middle Ages onward is heavily debated and probably impossible to know for sure--most references to them in medieval sources talk about people stealing stones to make them worthless for carts. Still, the evidence is pretty strongly in favor of roads to get people where they wanted to go.

The biggest argument in favor of roads is actually bridges. One of Boyer's most interesting observations is that Roman armies and settlers sometimes built bridges--sometimes quite lovely and sturdy stone ones, especially in Italy--but frequently they just forded rivers. Medieval people strongly preferred to build bridges.

It's possible there was superstition or folklore involved; you can definitely find scholars who've argued that. What is clear is that bridges were important sources of money from Merovingian times onward (...in Merovingian territory, which is Boyer's typical focus)--but they largely did not, it seems, make their money off locals, who typically got a free use pass. This is crucial, and it leads to one of the most important streams of continuity across early medieval dynasties and empires. The construction and maintenance of bridges and roads-Merovingian, Carolingian, Anglo-Saxon, take your pick-was almost always considered the people's responsibilities. Holy Roman Emperors get very cranky trying to pester their lords into pestering their peasants to maintain their bridges. Free use, we assume, was the benefit of being the ones to build and rebuild the thing in the first place.

Also in favor of roads is the geographic arrangement of settlements in early medieval Europe. As David Herlihy shows (see Medieval Households in particular), patterns of settlement in the early Middle Ages tended to be isolated clumps: a bunch of villages around some kind of central point, usually not yet a city. Although long-distance trade would have been mostly minimal (but see Michael McCormick's argument that human trafficking knit together the economies of the early medieval Mediterranean world into north and eastern Europe), local traffic--from village to castle, for example, or village to village--would have been predictable and repetitive. My understanding is that the road system was especially well developed in England, although I'm not sure whether that refers to more local roads like I'm talking about here, or the longer-haul ones like the roads that supposedly ran parallel to the Rhone in France, one on each side.