The exposure is based on the amount of light hitting the subject intertwined with how much light is reflected from the subject. Thus the exposure remains a constant regardless of camera to subject distance. While this might seem to violate the fact that light falls off with distances, it doesn’t because this is a special case.
Light falloff with distance is called the “law of the inverse square”. Suppose a lamp 1 meter from a surface delivers 1000 units of light. If we double the lamp to subject distance by backing off the lamp to 2 meters, the light falloff is 2 squared = 4. Now the light intensity at the subject plane is 1000 ÷ 4 = 250 units. But, you recognized this fact so what’s happening with our photo setup?.
The law of the inverse square only strictly applies only if the lamp is a point source like a tiny bare light bulb. As soon as we place this lamp in a reflector, or impose as diffuser, this law goes out the window. Maybe not completely gone, the degree violation is a variable, depending on the situation.
Suppose the lamp is put in a collimating reflector and the beams become parallel like a spot light? Now the spot does not obey, the falloff is practically nonexistent. Same for a laser beam, they practically never falloff, they can hit the moon with well-nigh no loss.
If the light bulb is in an umbrella and totally diffused, now the light is called a “broad” and this law goes out the window, you can move the subject around quite a bit and the exposure will be highly constant.
So what about a portrait subject illumined for an exposure of f/5.6? The light reflections from the face and clothing consist of highly diffused light beams. They don’t even come close to obeying the law of the inverse square. You move the camera all over the place and the exposure remains constant. However, just pat a bare bulb lamp and change lamp to subject distances and the exposure dances.
By the way, the popularity of the umbrella lighting and their origin, a broad, is due to the diffusion they bring to the table due to the fact they almost completely slay the law of the inverse square.
Added thoughts:
Spotlights output parallel beams. It is this parallelism that thwarts ray scattering thus the output of the spotlight is preserved over distance. Now most illuminated objects do not have polished surfaces thus they reflect light rays that scatter in all possible directions. Most of this reflected light from objects will be lost to us and our camera. If we draw trace lines of the light rays reaching our eyes and camera, the trace reveals, these image forming rays are arriving as parallel or nearly so. It is this parallelism that quashes the inverse square law. This explains why commonplace objects do not brighten or dim as distance changes and why we need not change camera setting as subject distance changes, and why spot light meter reading do not change with distances.