swiveled back in order to 
continue looking directly at 
the portraitist. If, however, 
while the person (not the 
picture of the person) held 
either of these directions of 
gaze, we were to view him or 
her from the side, the person 
would obviously no longer seem 
to be looking at us. But when we 
view the picture of the person 
from the side, the direction of 
the person’s gaze remains 
unchanged. In my opinion, this 
is the main reason for the 
curious effect. Of course, as we 
move to the side, the projection 
of the eye and head in the 
picture will become distorted 
and our perceptual system will 
correct for these distortions so 
that the shape of the eye and 
head will be perceived 
veridically. This step of 
correction is no doubt part of 
the explanation, but the effect 
will occur even without such 
correction. The perceived 
direction of the person’s gaze is 
not altered very much if at all 
even while other kinds of 
distortion occur from abnormal 
positions of viewing (e.g., the 
shape of the image of the entire 
eye changes).