Pictorial representations may be useful in ethics in a number of ways. In some cases it may be easier to convince someone that an act or state of affairs is wrong by pictorially representing it than by describing it with words. Think of how the video clips of Rodney King being beaten were so effectively used in court. One needs to wonder in these cases whether there is something that is in principle more effective about images than words, since prosecutors are now learning that having victims orally recount the crime before a jury can have a similar effect. Is there some sort of ``moral content'' that can be conveyed more efficiently with pictures than with words? Is there some moral content that can only be conveyed with images?
I'm told that interactive media have been used with some success in ethics courses. Students are presented with a pictorially represented ethical scenario in which they can intervene. They can literally watch the consequences of their intervention, and then draw their conclusions from there. In more abstract areas, it's certainly true that mere abstractness does not preclude pictorial representations. However, in the case of numbers and other abstract mathematical objects, it's slightly risky to describe these representations as pictorial. They are at least not pictorial in a realistic way, or so it would seem. I.e., they don't mirror the world any more than a sentence in a natural language mirrors the world (logical atomism aside.) These representations of numbers are useful because they allow us to literally see properties of numbers that are hidden in conventional symbolic form.
If it is true that non-propositional representations of abstract objects and states of affairs are no more or less realistic than linguistic representations, then the preferred medium of representation should be decided by which can convey more information, or convey information more accurately, or in more detail. Efficiency may be a criterion as well.
Perhaps diagrams and other pictorial displays would be more helpful in areas like mereology than sentences ever could be. As teachers, we all know that some people are much more responsive to diagrams than linguistic descriptions. Perhaps in some areas we are all more responsive to pictures than words.
I suspect that some of these questions will begin to be answered as we learn more about how our minds/brains actually process information (in particular, questions about the suitability of certain media for conveying ideas.) But this research would seem to leave untouched basic questions about how various kinds of non-mental representations represent their objects.