The horizontal shear stress you describe is caused by the need for the top and bottom of your joist to maintain their lengths in a bending situation. If you take a stack of paper, hold it horizontally and bend it in the middle, you will notice the ends of the stack remain vertical, and the individual sheets of paper have slid among each other to accomplish this. You can further understand this concept by drawing several evenly spaced vertical lines along the edge of the stack. Now when you bend the stack, you will notice the lines remain parallel to each other, and vertical, but not perpendicular to the stack. Assuming you could create a parabola with the stack, you would notice that in the center of the stack, (the apex of the curve) there is absolutely no shear at all. The line in the center is perpendicular to the stack no matter how far you bend it. However, the farther away from the center, the more the lines diverge from perpendicular, ie the more "horizontal shear" there is. There is a major difference between a stack of paper and a joist however, and that is that a joist's fibers do not "slide" among one another. In your stacked 2x4 example, if you were to place two nails at the exact center of your span, there would be no net change in their strength. If you place one nail at each end, the change is dramatic. I would never recommend a series of holes along the center of a joist along its entire length, for the reason you illustrate. However, the shear force you describe applies to the entire vertical component of a joist. There is no difference between the top or the bottom.
My comment that I wouldn't worry about the top of the joist in this particular case is for several reasons: 1) wood fibers are much stronger in compression than they are in tension. 2) an inch and a half of material (at the point of the hole) is sufficient to maintain integrity (look at any engineered truss joist) 3) The subfloor above this joist adds at least another half inch of material (x the area of the floor itself)
So, no. I haven't changed my position. The lateral forces on a joist are greatest the farthest away from the center, as are the vertical forces and the horizontal and vertical shear issues.
Horizontal shear is uniformly applied to the vertical component of the joist, increasing in value from the center to the ends of the joist.
Again, all of this is mute, because the situation originally described is a static one that has been that way for 20 years. It's not going anywhere, but if it makes him sleep better at night, he can add the strap or build a wall downstairs.
Keep asking questions, maybe we can come to an agreement someday...