Abstract: Extensive polygonal networks of normal faults have reportedly been identified within layer-bound sequences in about 28 sedimentary basins worldwide. The gentle regional dips, passive tectonic settings and geometry of the fault networks have led to the conclusion that faulting must have resulted from gravity-driven mechanical compaction. The faulted sequences comprise very fine-grained sediments, with lithofacies that range from smectitic claystones to almost pure chalks. In most, if not all, cases it is clear that volumetric contraction has occurred with horizontal contraction of the sediments complementing the heave of the faults. One explanation which has previously been offered is that the fine-grained sediments have shrunk due to syneresis, a process that involves spontaneous contraction of the solid network with expulsion of the pore fluid. However, syneresis is an implausible mechanism because it does not explain the observed lithological variation in the sediments concerned, why the initiation of faulting occurs in the depth range 100–1000 m, and why faulting continues for millions of years. A much simpler explanation is that shear failure inevitably results from one-dimensional compaction if the coefficient of friction is sufficiently low; and there is some evidence from laboratory measurements that the coefficient of friction is likely to be exceptionally low in these fine-grained sediments. Qualitatively, low coefficients of friction also explain why these compaction faults preferentially dip towards the basin margin where the regional dip of the bedding is greater than 1˚. Furthermore, they help to explain the origin of a polygonal fault system in the Eromanga Basin, South Australia, where the situation is complicated by the presence of a low velocity, ductile layer at the base of the faulted sequence.