Brian,

the rotation of the vectors of North America between the Wasatch front in Utah and the San Gregorio fault (say, at latitude of Santa Cruz City California) is indeed largely due to deformation, but there would also be an effect of spherical rotation. I don't know whether the effect of rotatation is in the same direction as effect of deformation or not.

All plate rotates about a pole on a more or less spherical earth. Some poles are close to the plate and some are on the other side of the earth, depending on the plate and where you are.

Continental plates away from the cratons (the oldest part of the continents) need not be rigid. Continents are weaker than oceans because their rock contains quartz and other minerals that start to flow at a lower temperature than oceanics crust minerals (or mantle lithosphere). That is why seismicity in California south of Mendocino is generally all above 20 km depth: it creeps below that.

I don't know how much strain can be stored elastically within oceanic plates. Lets say they are squeezed and store 40 m of strain. But they flex only at the subduction zones. If there is a subduction quake that releases 20 m of this, it is released close to the subduction zone. The remaining 20 m does not do anything, it is still there, and does not get recorded as a displacement.

The above argument is just logic based on some of the below.

Before the global GPS and VLBI geodetics, the only way to get the Pacific-North American plate motions was to do global plate circuits. The most relevant papers are Stock and Molnar 1988 and Atwater and Stock 1998. Global Plate Circuits use the magnetic reversal pattern in the oceanic crust along with age of those reversals to get velocities at all different places, poles of rotation, etc. You can't do this across subduction or strike-slip boundaries. So, you go across spreading centers something like North-America-African, African-Indian, Indian-Antarctic, to Antarctic-Pacific.

From that you get Pacific North American. The exercise would make no sense if the oceanic plates were deforming. So, yes, that rigid.

For the last few million years only there has been spreading between Pacific and North American plates with interpretable magnetic anomalies in the southern Gulf of California. De Mets and others published a paper that came up with about 48 mm/yr (a bit more complicated than that). But they missed motion west of Baja, and they underestimated the motion by about 4 mm/yr.

Chris