How it normally works with a lens with an aperture control ring: the camera will meter with the lens wide-open, and when you press the shutter, the lens will stop down until it is physically stopped by the setting of the aperture control ring. The camera will know ahead of time what the aperture ring's setting is because there is a position lever on the camera that is engaged by and controlled by the meter coupling ridge on the lens's aperture ring. This allows the camera to use the setting to calculate aperture-priority autoexposure.

However, in the case of G lenses that don't have aperture rings, the only physical limit will be the smallest aperture of the lens — ƒ/16. Additionally, because the lens does not have a meter coupling ridge on the aperture control ring to push against the camera's aperture position sensing lever, that lever is left in the state that would indicate the widest exposure setting possible. As a result, the camera will always meter what it thinks is wide-open (ƒ/1.8), but the lens will always close down to ƒ/16 when the shutter is pressed. So unless you use anything but full manual exposure control (where your calculations assume ƒ/16), your images will come out very underexposed (by over 6 stops).

See this dpreview.com discussion: G lens on an old SLR.