What is this lady looking at? If your answer is "herself in the mirror," you are among the majority, but you are wrong. This is an optical trick used from medieval painters to modern directors. It's the need to subvert physics for artistic effect. Discover the truth about the Venus Effect.
There tend to be a lot of pictures of Venus in various eras of art, mostly because the goddess of love allowed both patrons and artists to explore their very favorite subject — naked women. Venus, in mythology, was a great beauty who was unhampered (as some of her fellow goddesses were) by great modesty. This allowed painters to show Venus unclothed (and is the reason why the Rubens image at the top of this entry had to be thoroughly cropped). Venus was also famously vain, and so showing her staring into a mirror makes perfect sense. If you are an artist, it allows you to show multiple angles of the subject. Except that it doesn't. Although most descriptions of these paintings claim that Venus is looking at herself in the mirror, any time spent actually thinking about the angles involved means she couldn't possibly be seeing herself with the glass held at that angle. Venus is not looking at herself. Venus is looking at the painter — or the viewer. The fact that people assume she's looking at herself is due to what's known as the Venus Effect.
Why do people assume that Venus is looking at herself? Maybe it's in part because people have heard enough stories about what Greek deities did to people who spied on them when they were nude. Artemis famously turned one man into a stag and had his own hounds hunt him down and kill him. On the other hand, it might be a more fundamental misunderstanding of physics. One group of researchers, from the University of Liverpool and the University of Manchester, found that the conclusion that the woman in the picture is staring at herself in the mirror shows a fundamental shortcoming in the way people think about reflections and mirrors generally.
The researchers showed people old paintings, but also photographs of people near or reflected in mirrors, and even asked people to sketch out the sight lines into mirrors. They played with ways to orient the subject in the photographs so that the Venus Effect would be completely eliminated. Overall, they had very little luck. The research subjects stuck to the interpretation that what they could see, the subject of the picture could see as well, even if they knew the photographs were a little off.
Regardless of the position of the mirror or the subject in the pictures, people tended to think that the subject of the picture could see themselves if they appeared anywhere near a mirror. The Venus Effect kicked in even more dramatically when the subject of the picture was — to the observer's eyes — reflected in the mirror. This happened when any part of them was reflected, not necessarily their face. It seems that people aren't able to put themselves in the position of the subjects of the paintings.
Possibly, though, the Venus Effect is about something else. People might be paying more attention to the narrative of the image than the physics of the image. We see this a lot in movies, when some character appears to be staring moodily into a mirror and the audience sees their expression as they look at themselves. Often, we're actually seeing their expression as they look at a camera. The angles don't work for the character to be staring at their own image. Of course we know that we aren't meant to be thinking of that.
The scene we're meant to see is that character's reaction to their image and, in order to get the overall visual picture right, the audience sees an image that the actor doesn't. Paintings, especially old ones, were no less a narrative than a movie. Some of these early Venus Effects might be a nod to Venus looking out at the observer. In other paintings, such as the Marie Cassat picture of the mother and child, the scene the painter depicting doesn't make sense unless the kid is actually looking at herself. Otherwise, why would they be holding up the mirror? The painting is a little scene, and the need to make the scene make sense trumps the need to get the angles right. And so, in photos, we assume that people are looking at themselves, even if they can't be. The Venus Effect might not be a defect of geometry, but an understanding that because something is being photographed, is meant to be communicating a message to us, and not drilling us on our spatial reasoning.