How the Vikings Crossed an Empty Ocean

AThe Norse voyages to Iceland and Greenland are usually explained by pointing at objects. A notched wooden disc, a lump of transparent crystal, a bearing dial: each has been offered, at one time or another, as the device that turned an eight-hundred-kilometre passage out of sight of land into a routine act rather than a gamble. The appeal of such explanations is partly practical, since an object can be photographed, reconstructed and put in a case, whereas a habit of mind cannot. It is also, however, an argument from absence. The medieval annals that record these crossings describe cargoes, quarrels and losses at length, yet mention no instrument at all, and a catalogue of grave goods published in 1998 lists nothing from the several hundred burials of seafaring men in the northern settlements that a modern sailor would recognise as navigational equipment. Wood decays and crystal is easily mislaid, and nobody expects a complete record of what was carried in an open boat. Historians have generally treated the silence as an accident of preservation. A smaller group has begun to treat it as evidence.

BThe most developed version of that second position belongs to Astrid Villard, a maritime historian at the Kvitsoy Institute for Northern Seafaring, who argues that the decisive equipment was never carried aboard but held in the memory of the crew. Norse sailing directions, preserved in later manuscripts, are not charts but sequences of expectations: the number of days to be spent on a heading, the bird that ought to appear, the colour the water takes over a particular bank, the height at which a mountain should stand above the horizon when land is finally raised. Villard assembled 214 crossings named in the annals of the northern dioceses and found that 27 of them ended in loss, a proportion that fell to one in fourteen on the Iceland run, which was sailed most often of all. Her explanation is not that those sailors were lucky. A route travelled every season, she maintains, is a route under continual correction, since every returning ship brings back a revision, and the description held by a navigator is amended at a speed no written chart of the period could approach.

CVillard is careful about how far the argument reaches. Knowledge of the kind she describes decays quickly, because it exists only in the heads of the people who use it, so a route left unsailed for a generation is in effect a route that has to be found all over again. That, on her account, is why the western passages beyond Greenland were made a handful of times and then let go, while the Iceland run stayed ordinary for three centuries. Two weaknesses in her evidence are acknowledged. The annals she counted were compiled long after the events, mostly by clerics with no interest in seamanship, and a voyage that ended badly was far likelier to be written down than one that ended in the usual way, which would push her figure for losses upward rather than down. She also has no means of dating the sailing directions themselves, since the manuscripts are late, and a remark about a bird or a swell may have entered the text at any point across three hundred years.

DDiarmuid Callaghan, an experimental archaeologist at the Ardmore Centre for Maritime Reconstruction, dismisses the whole approach as an argument that cannot fail. Expectations about birds and water colour, he objects, tell a crew where they are not; they do not hold a vessel on a line between one noon and the next, and it is the accumulation of small errors in heading that leaves a ship a hundred kilometres south of an island it never sees. Between 2011 and 2016 his group sailed a replica trading vessel across the same waters eleven times, recording position continuously. Crews steering by sun and swell alone wandered an average of 9 degrees from the intended course each day, an error that a shadow board of the type reconstructed from the Greenland disc brought down to 3 degrees. Callaghan concedes that his crews were modern, that every voyage was made between June and August, and that a single hull was used throughout. What he denies is that these limitations touch his central point, since error accumulates faster, not slower, in the conditions he avoided.

EMost specialists now describe the two accounts as answering different halves of one problem, and that division of labour has become the standard settlement: a device of some sort keeps a ship pointing in roughly the right direction, while remembered expectations tell the crew when the direction has become wrong and what should be done about it. Neither half is much use without the other. The settlement, though, conceals a genuine dispute about what the Norse achievement consisted of. Callaghan reads it as a technical solution, arrived at once and thereafter passed on; Villard reads it as a social one, sustained only for as long as enough ships kept sailing to keep the knowledge current, and lost when they stopped. The distinction matters well beyond the north Atlantic, since it decides what a seafaring society must maintain in order to remain capable, and on that question the archaeology of the seabed, which recovers hulls and cargo but never conversation, will always speak for one side more loudly than for the other. Whether the capability was a possession that could be handed on, or a practice that had to be performed, is in the end what the objects cannot settle.