This is not quite right.
By specifying print size, original size (FF sensor) and viewing distance you have *fixed* the CoC and as you indicate it is usually based on typical human vision. This is the size of an object in the original image that is perceived as a "point" in the print at the specified viewing distance. Anything object bigger than this size in the roiginal image is perceived as something bigger than a "point", or smallest resolvable detail, in the viewed print.
In your examples, the CoC is around 35µm, and something like this was used by lens makers to make DoF scales, based on a ~8x10-12" print viewed from about a foot. (Similar to your 2'x3' print viewed from 3'). This also goes along with "300ppi as a guideline for maximum useful detail". So the CoC is larger than the Airy disk and the pixels for both cameras, so they do not play a role in your scenario. If you use pixel size to define CoC, then you are no longer viewing the 24x36 print from 1 m. The key parameter in your discussion should be the size of the original pixel and how much it has been enlarged to the print (simply put, how many original pixels are in an inch of print). If those are printed big enough that they represent the CoC, you have a much bigger print from the 50MP camera than from the 18 MP camera. In your examples, the 18MP image is printed at ~150ppi and the 50MP one is ~240ppi. So the 50MP will appear more detailed, simply due to the increased resolution If you print both at 150 ppi, you get a ~24x36 for the 18 MP image and ~36x54 print from the 50MP. (You are right that it is multivariate When you compare these two prints, do you use the same viewing distance?) The large print, however, has been enlarged more from the original FF sensor dimensions so that all types of blur (OOF, motion and diffraction) have become more visible. The effects of diffraction only become more visible if you blow up the image more. As you noted the amount of diffraction (Airy disk size) is the same, regardless of the pixel size and is completed before the image hits the sensor. The higher MP sensor just does a better job at capturing the details and captures a higher resolution version of the same Airy disk.
Simply put, for prints of the same size, smaller pixels will never yield a worse image than big pixels and the effect of diffraction is the same for both prints. If you use all those nice pixels to allow a bigger print at 300 ppi, then diffraction will be more noticeable because of the extra enlargement and the likelihood of closer viewing distances