[1] Among the most puzzling artifacts in the history of navigation are the portolan charts, the sea maps of the medieval Mediterranean whose earliest surviving examples, from the late thirteenth century, depict the coastline with an accuracy that would not be matched by land surveys for centuries. [2] Crisscrossed by radiating lines of bearing that fan out from compass roses, the charts were plainly working documents, creased and stained by use, yet they appear in the record already fully formed, without the halting predecessors that usually mark a craft's development. [3] How mariners lacking any systematic means of fixing longitude produced coastlines so nearly correct remains contested.
[4] One school holds that the charts were compiled empirically, stitched together over generations from countless recorded compass bearings and estimates of distance sailed between headlands, the errors of individual voyages canceling one another in the aggregate. [5] A rival account proposes that the accuracy is too great for such accretion and points to a lost tradition of systematic measurement, perhaps inherited from antiquity, whose records have not survived. [6] A third position sidesteps the question of origin and observes that the very grid of bearing lines could have functioned as a self-correcting framework: a chartmaker copying and recombining existing charts against fresh reports would tend, over time, to reconcile discrepancies and tighten the fit. [7] On this view the charts' precision is neither a sudden inheritance nor a simple sum of observations but an emergent property of the copying tradition itself. [8] The scarcity of early examples, likely worn out at sea, leaves each hypothesis underdetermined by the evidence, and the debate turns as much on what one expects a craft to leave behind as on the charts themselves.
