The Lenore Thomson Exegesis Wiki
Perceivers make lenses for themselves through which to view the world, and then they look at everything they can. It doesn't matter if the lenses are imperfect, or if you are pointing them in the wrong direction. As long as you can see something through your lens, you are doing fine.
INJs collect and carefully index all the lenses that are ever made, but never look through them. You can show an INJ any image, and he will be able to show you which lens was used to create it. Yet the concept of looking at anything through a lens himself would be alien to him.
The goal of the Pe worldview is to construct one perfect lens to allow you to view the world subjectively (through Ji). The goal of the Ni worldview is to collect all the lenses that anyone ever looks through. By collecting all the lenses, you aren't confined to the imperfections inherent in any one lens. Thus, you can develop a basis for viewing the world objectively through Je.
The IP worldview is like setting up a still life and photographing it or painting it. You arrange the objects in a harmonious way, you set up lighting that makes them look good, and choose the best possible angle (Pe). Then you paint as completely as possible from that angle by using perspective to project the 3-D objects onto a 2-D space (Ji). In other words, you try to capture reality by taking a subjective, temporal perspective, and making your depiction as complete and believable as possible. You find truth by manipulating the context (positioning, lighting, view angle, etc.), and you use this context as a vehicle for your image.
In fact, perspective projection is an excellent example of the way Ti brings coherence to reality. Ti involves making subjective choices about how to weigh the importance of every element that composes the gestalt. If something is close to you, you paint it larger, even though in reality, objects don't grow larger as they get closer to you. Another example is the way that picking a certain view angle and setting up the objects in a certain way means that some objects will occlude others. If the orange is in front of the vase of flowers, then you don't draw part of the vase because it is hidden. With Ti, you must make that kind of choice to better convey the harmony inherent in the objects you are depicting, even though you willingly sacrifice a bit of data to do so (specifically, the part of the vase that is hidden). You don't even bother trying to paint them as they "really are," because it is impossible to fit 3-D objects onto a 2-D space. Instead, you paint them as they appear, so that people can figure out how they really are through induction. It doesn't matter if the orange is hiding part of the vase, as long as people can guess what the rest of the vase looks like.
From the standpoint of Ni, this process is unacceptable. Arranging the objects in advance? How arbitrary and contrived. Setting up lighting? That doesn't tell us anything about how the objects really look, because another lighting set up would make them appear differently. Perspective projection? Completely arbitrary, because it distorts objects based on your current viewing position (which is also arbitrary and means that you forgo viewing the objects from another position). Objects occluding others? Useless, because then hidden aspects of the objects are never depicted.
Thus, cubism and modern art are heavily influenced by Ni because they seek other ways of seeing the world. For instance, Picasso paints figures that appear to be in multiple positions at the same time. Another example is the "Cine-Eye" of Russian filmmaker Dziga Vertov. Ni art seeks to disentangle objects from their position and temporal contexts.
The ultimate Se representation of an object is a photograph, because it captures the way the object impacts your senses during a certain instant. The ultimate Ni representation of an object would have to show the object at every possible position at every moment in time. In the end, NJ representations tend to regress into symbolism, which disentangles an object from the context that is created by viewing it, and replaces it with something that people can agree on to represent the object.