That taper on the upper part of the shank is not corrosion eroding away the lower part of the shank. It is the primary design feature of the Greer patent according to the book,
Railroad Spikes by James M. Joyce. According to the book, the main claim for the Greer spike is its extra holding power from what it refers to as the unusual swelling shape near the top of the shank. A separate driving head is also claimed as a feature, but there is no explanation of why the separate head would be an advantage.
However, according to the book, the Greer spike had a disadvantage in that it easily bent outward by the thrust of the rail base. Apparently this bend occurred at the point where the wider part of the shank met the narrow part. The discovery of this defect quickly curtailed the use of the Greer spike on the RGS.
From the illustration of the Greer spike, it appears as though all of its profile were cut in only two dimensions. In other words, if the spike is say ¾” wide (in the direction of the rail length), it is shaped as if it were cut out of a ¾” thick steel plate. However, I am not suggesting it was made that way. I suspect that it was made by rolling the profile shape as if in rolling a rail, and then the individual spikes were somehow sheared or sawed off the long rolled shaped member. This left all of the edges of the spike sharp. The gentle rounded edges appearing in the photo posted by Russo Loco apparently are due to corrosion material loss.
The Greer spike was also used on the Kansas City, Fort Scott & Memphis RR and on the Silverton RR.