The comparison is with the viewing angle we tend to look at pictures with. The idea is that when we look at a picture taken with a 'normal' lens, the taking angle will roughly equal the viewing angle. This means that the perspective will be natural.
If the viewing angle is less than the taking angle, the perspective will appear to be expanded. If the viewing angle is greater than the taking angle, the perspective will appear to be compressed.
The 'normal' focal length is generally assumed to be roughly equal to the diagonal of the film or sensor format. It isn't an exact thing, of course. This is based on the idea that we look at pictures from a distance that is approximately equal to the diagonal of the picture.
Best,
Helen