Well, given a step wedge and an image on the same film and processed the same way, any fog is present in both and is "used" to produce any image from the negative. So, what you measure and see are the same. The difference is that your eye integrates the entire image whereas the densitometer does not.
The integration is the major difference between what you see and what you measure, and not the fog.
Your eye sees the entire picture and judges contrast, the densitometer sees one step at a time and measures contrast. These two pieces of data do not need to correspond. Thus, fog is different to you as the observer and the curve that might be drawn.
In making enlargements or contact prints, fog tends to lower contrast even though it is uniform throughout. That is because it lowers the difference between Dmax and Dmin. However, by use of the right print contrast, this can be removed.
Think of it in terms of the signal to noise (S/N) model used in electronics. The fog is the 'noise' and the negative density represents the 'signal + noise'. When the 'signal' value is low, the 'noise' value becomes significant; when the 'signal' is high, the noise is insignificant. The result of the fog ('noise') in photography is to lower the image contrast (as Ron has said).
I thought I would revive this thread with a somewhat cryptic note in Haist's sensitometry chapter regarding D-min (fog) in relation to the plotting of a characteristic curve, and what made me start to think about this and theoretical implications to ISO testing etc.
"In some cases the D-min has been subtracted from all densities, making the D-log E (H) curve touch the line of zero density."
As we know this is a common practice when people perform various types of Zone System tests for personal EI and contrast. It is advocated by Adams etc. The next sentence in Haist is:
"This practice, however, fails to take into account that fog silver is not distributed uniformly over all levels of image density."
Nothing more is said, but it got me thinking. The footnotes to this sentence reference two sources which may or may not have any relevance today:
R. B. Wilsey, "Fog Corrections in Photographic Densities," Phot. J., 65: 454 (1925)
H. A. Pritchard, "The Fog Correction of Photographic Densities: A Sensitometric Study," Phot. J., 67: 447 (1927)