Models for the tectonic role of the left-slip Garlock fault in the context of the western North America right-lateral transform plate boundary include: (1) a conjugate to the right-slip San Andreas fault, (2) an intracontinental transform structure facilitating Basin and Range extension to the north, and (3) a clockwise-rotation structure. Here I suggest that the Garlock fault experienced clockwise rotation (~4°/myr), driven by right-lateral shear across the western North America plate boundary, as part of a non-rigid bookshelf fault system. A classic rigid bookshelf-fault model requires space issues to arise at the bookshelf-panel ends, which can be accommodated by non-rigid off-fault deformation that ultimately leads to an observed slip gradient with highest slip along the central fault segment and negligible slip at the fault tips. My kinematic model explains several peculiar observations about the Garlock fault, including: (1) a bidirectional decrease in fault slip rate and magnitude from the fault center (i.e., 65 km to <15 km and ~10 mm/yr to ~1 mm/yr, respectively), (2) alternating domains of extension and contraction in the fault-termination quadrants (e.g. extension in the Antelope Valley and Basin-and-Range quadrants and shortening in the Avawatz and Tehachapi Mountains), and (3) maximum strike-slip offset of ~65 km along the central fault segment, assuming 40º clockwise rotation since 10 Ma. Paleomagnetic tests of the required vertical-axis rotation are inconclusive, with both clockwise and counter-clockwise rotations. However, counter-clockwise observations are from sites located adjacent to the fault, and probably record drag effects associated with left-lateral shear. This model predicts that slip rates and seismic hazard are highest along the central Garlock fault. I use simple geometric arguments to test whether earthquakes on the Garlock are guided by activity along the southern San Andreas fault, which show that right-lateral shear along the San Andreas predicts the observed magnitude of bookshelf offset along the Garlock on overlapping timescales. This provides further evidence that loading of the Garlock fault is temporally variable and that its earthquake cycle is closely related with the San Andreas fault.